KWAME NKRUMAH UNIVERSITY OF SCIENCE AND TECHNOLOGY, KUMASI – GHANA INSTITUTE OF DISTANCE LEARNING [] MATH 251: ENGINEERING MATHEMATICS IV (CREDITS: 4) YAO ELIKEM AYEKPLE 1 Publisher Information © IDL, 2008 All rights reserved. No part of this book may be reproduced or utilized in any form or by any means, electronic or mechanical, including photocopying, recording or by any information storage and retrieval system, without the permission from the copyright holders. For any information contact: Dean Institute of Distance Learning New Library Building Kwame Nkrumah University of Science and Technology Kumasi, Ghana Phone: +233-51-60013 +233-51-61287 +233-51-60023 Fax: +233-51-60014 E-mail: idldean@kvcit.org idloffice@kvcit.org kvcit@idl-knust.edu.gh kvcitavu@yahoo.ca Web: www.idl-knust.edu.gh www.kvcit.org 2 Publisher notes to the Learners: 1. Icons The following icons have been used to give readers a quick access to where similar information may be found in the text of this course material. Writer may use them as and when necessary in their writing. Facilitator and learners should take note of them. Icon #1 Icon #2 Learning Objective Activity Icon #6 Icon #7 Time For Activity Icon #11 Well Done Self Assessment Icon #12 Icon #3 Icon #4 Introduction Information Icon #8 Icon #9 Group Discussion Icon #13 Read Icon #14 Note/Learning Tip Pause Icon #5 Summary Icon #10 New Terms Icon #15 Online Question 2. Guidelines for making use of e-learning support This course material is also available online at the virtual classroom (v-classroom) Learning Management System. You may access it at www.kvcit.org 3 Course Writer Yao Elikem Ayekple received both BSc. (Hons) and MSc. Mathematics from the Department of Mathematics, Faculty of Science, KNUST, Kumasi. He entered the university in 1997 for undergraduate degree and completed in 2001. After National Service in 2002, he enrolled for his MSc. Mathematics programme. He completed his MSc. Thesis in 2004 and defended in the same year. He was a demonstrator in the university from 2002 to 2004. He was employed as a Teaching Assistant from 2004 to 2005. From 2005 to date, he has been teaching at his alma mater, Kwame Nkrumah University of Science and Technology, Kumasi at the Department of Mathematics as a Lecturer. 4 Course Introduction Differential Equation is among the lynchpins of modern mathematics, which along with matrices, are essential for analyzing and solving complex problems in engineering, the natural sciences, economics, actuarial sciences, and even business. In the first section, we begin with general differential equation and classifying them in terms of type, order, degree and linearity. We then narrowed to one type of differential equation, which is ordinary differential equation. We move on to look at first order ordinary differential equation types, put them in classes and find an appropriate methods of solving most of them, if not all, and its applications to practical problems. We then consider higher order linear ordinary differential equation, that is, orders greater than one. Initially, we concentrate on second – order homogeneous linear ordinary differential equation with constant coefficient because their theory is typical of that of linear differential equations of any order n (but involves much simpler formulas), so that the transition to higher order n needs only very few new ideas. The Wronskian is used to determine the independences of solutions to a particular equation. Elementary methods such as undetermined coefficient, variation of parameter and linear differential operator are used to solve second – order non – homogeneous linear ordinary differential equations with constant coefficient. Systems of differential equations occur in various applications. Linear systems are best treated by the use of matrices and vectors, of which, however, only modest knowledge will be needed here. We will deal with first – order linear systems with constant coefficient and draw some analogies from the first order ordinary differential equation. 5 Course Objectives On completion of the course, it is hoped that you would be able to: Classify first order ordinary differential equations in terms of type, order, degree and linearity. - Have the technique to solve most, if not all, first order differential equations - Be able to solve differential equation of higher order with constant coefficient using the appropriate technique for each kind of higher order equations. - Apply the technique in solving the differential equation in solving differential equation appears in actuarial science, economics and finance. - Reduce higher order linear differential equations to system of first order differential equations and solve. - Transform system of equation to normal form and put it in matrix and vector to solve using eigenvalues and eigenvectors. Course Outline Unit 1 Basic Concepts: o Definition of Differential Equation o Classification of Differential Equations o Solution of a Differential Equation o General Solution o Initial – Value and Boundary – Value Problems Unit 2 Ordinary Differential Equation of First Order: o General and Particular Solution of a First – Order Differential Equation. Method Separation of Variables Reduction to Separable Form Exact Equation Integrating Factors Linear Equations and Those Reducible to that form Linear Equations Reducible to Linear Forms by change of Variable Bernoulli’s Equation 6 Riccati’s Equation o Applications of First – Order Ordinary Differential Equation Unit 3 Linear Differential Equation of Higher Order: o Homogenous Equation with constant coefficient o Theory of solution of Linear Differential Equation Independent and Dependent function Wronskian o Linear Non – Homogeneous with constant coefficients Method of Undetermined Coefficient Variation of Parameter Linear Differential Operator Unit 4 LAPLACE TRNSFORM: o Definition of Laplace Transform o Find Laplace Inverse o D The Differential Operator Reduction of n th Order Equation to a System of First – Order Equations. The Matrix Method The Eigenvalue Method of Homogeneous Linear System Non – Homogeneous Linear System. Mode of Assessment This would involve a combination of continuous assessment (30%) and end of semester examination (70%). The continuous assessment would include handed in exercises and midsemester examination. Test your understanding Within each unit/session are exercises. These are meant to help you assess your understanding of the unit or course. It is vital that you take time to complete or find solutions to these exercises as they occur in the study material. Make sure you go through them. I recommend you have a notebook/exercise book specifically for this purpose and keep it with your study materials as record of your work. 7 References 1. Frank Ayers, JR.: Theory and Problems of Differential Equations, McGraw – Hill Inc. 2. Earl D. Rainville and Philip E. Bedient: Elementary Differential Equations, The Macmillan Company 3. Richard Bronson and Gabriel Costa (2006):Differential Equation (Third Edition), McGraw – Hill Inc. 4. Erwin Kreyszig(2002): Advanced Engineering Mathematics (Eighth Edition), John Wiley and Sons, Inc. 5. William E. Boyce and Richard C. DiPrima: Elementary Differential Equations and Boundary Value Problems, John Wiley and Sons, Inc. 6. www.sosmath.com 7. https://tutorial.math.lamar.edu/pauldawkins 8 UNIT 1 BASIC CONCEPTS Introduction The laws of the universe are written largely in the language of mathematics. Algebra is sufficient to solve many static problems, but the most interesting naturally phenomena involve change and are best described by equations that relate changing quantities. Many important and significant problems in engineering, the physical sciences, and the social sciences such as economics and business when formulated in mathematical terms require the determination of a function satisfying an equation containing the derivatives of unknown function. Such equations are called differential equation. Perhaps the most familiar example is Newton’s law: d 2x m 2 F dt for the position (1) x t of a particle acted on by a force F . In general F will be a function of time t , the position x , and the velocity dx / dt . To determine the motion of a particle acted on by a given force F it is necessary to find a function x t satisfying Eq. (1). If the force is that due to gravity, then F mg and d 2x m 2 mg dt (2) On integrating Eq. (2) we have: dx gt c1 dt x t 12 gt 2 c1t c2 where c1 and c2 are constants. To determine (3) x t completely it is necessary to specify two additional conditions, such as the position and velocity of the particle at some instant of time. These conditions can be used to determine the constants c1 and c2 . This unit introduces us to the differential equation in general and helps to differentiate between various kinds and associated solutions. 9 Learning Objectives After going through this section session, you would be able to: • know what is a differential equation, and state the difference between the independent and dependent variables. • classify differential equations in terms of types, order, degree and linearity. • differentiate between general and particular solution. • derive differential equation from general solutions that is; elimination of arbitrary constant. • differentiate between boundary and initial conditions and be able to solve problems with such conditions. Session 1-1 Differential Equations The laws of the universe are written largely in the language of mathematics. Algebra is sufficient to solve many static problems, but the most interesting naturally phenomena involve change and are best described by equations that relate changing quantities. Many important and significant problems in engineering, the physical sciences, and the social sciences such as economics and business when formulated in mathematical terms require the determination of a function satisfying an equation containing the derivatives of unknown function. Such equations are called differential equation. Perhaps the most familiar example is Newton’s law: d 2x m 2 F dt (1) d 2x m 2 mg dt (2) x t of a particle acted on by a force F . In general F will be a function of time t , the position x , and the velocity dx / dt . To determine the motion of a particle acted on by a given force F it is necessary to find a function x t satisfying Eq. (1). If the force is that due to gravity, then F mg and for the position On integrating Eq. (2) we have: dx gt c1 dt x t 12 gt 2 c1t c2 where c1 and c2 are constants. To determine (3) x t completely it is necessary to specify two additional conditions, such as the position and velocity of the particle at some instant of time. These conditions can be used to determine the constants c1 and c2 . 10 Differential equations are an important part of the calculus, the fundamentals of which are presented here. In developing the theory of differential equations in a systematic manner it is helpful to classify different types of equations. There are two types of derivatives which usually are (and always can be) interpreted as rates. For example, the ordinary derivative dy / dx is the rate of change of y with respect to x (independent variable), and the partial derivative u / x is the rate of change of u with respect to x when all independent variables except x are given fixed values. And the study of differential equations has two principal goals: 1. 2. To discover differential equation that describes a physical situation; To find the appropriate solution of that equation. The solution of differential equations plays an important role in the study of the motions of heavenly bodies such as planets, moons, and artificial satellites. 1-1.1 Definition of Differential Equation A differential equation is an equation which involves one or more derivatives, or differentials of an unknown function. Example 1.1 The following are examples of differential equations involving their respective unknown functions: dy 1.1 3x 2 dx dR 1.2 kR 0 dt d2y 2 k y 0 dx2 1.3 x2 y2 dx 2xydy 0 1.4 2v 2v v h2 x2 y 2 t 1.5 L d 2i dt 2 R di 1 i E cos t dt C 1.6 2v 2v 0 x2 y 2 1.7 11 3 d 2w dw 2 xw w 0 dx dx 1.8 d 3x dx x 4 xy 0 dy dy3 d2y m dy 1 2 dx H dx 1.9 2 1.10 3 3 2 d y dy x e dx3 dx 3 1.11 d2y g sin y 0 dx2 l 3 1.12 d2y dy 3 dy 2 3 y y 5x dx dx dx 7 2 1.13 When an equation involves one or more derivatives with respect to a particular variable, that variable is called an independent variable. A variable is called dependent if a derivative of that variable occurs. In the equation: d 2i di 1 1.6 R i E cos t dt C dt 2 i is the dependent variable, t the independent varible, and L , R , C , E , and parameters. The equation : L 2v 2v 0 x2 y 2 are called 1.7 has one dependent variable v , and two independent variables x and y . Since the equation x2 y2 dx 2xydy 0 may be written x2 y 2 2 xy or dy 0 dx x2 y2 dydx 2xy 0 we may consider either variable to be dependent, the other being the independent one Oral Exercise 1 12 Identify the independent variables, the dependent variables, and the parameters in the equations given as examples in this section. 1-1.2 Classification of Differential Equations Recall that a differential equation is an equation (has an equal sign, but not an identity) that involves derivatives. Just as biologists have a classification system for life, mathematicians have a classification system for differential equations. Differential equations are classified as to: (a) type (namely, ordinary or partial) (b) order (c) degree, and (d) linearity (a) Type We can place all differential equations into two types: ordinary differential equation and partial differential equations. An Ordinary Differential Equation, (ODE) is one containing ordinary derivatives of one or more unknown functions (dependent variable) with respect to a single independent variable. Examples are equations: (1.1), (1.2), (1.3), (1.4), (1.6), (1.8), (1.9), (1.10), (1.11), (1.12) and (1.13). So also is each of the two equations with more than one dependent variable: dx dy (1.14) a b c dt dt dx dy (1.15) d e f dt dt A Partial Differential Equation is one involving partial derivatives of one or more dependent variables with respect to one or more of the independent variables. Examples are equation: (1.5) and (1.7). Note: The systematic treatment of partial differential equation lies beyond the scope of this book. We treat only ordinary differential equation with constant coefficient in this book. 13 (b) Order Definition 1.1 The order of a differential equation is the order of the highest-ordered derivative appearing in the equation. For instance, 3 d2y dy 2b y 0 2 dx dx (1.16) is an equation of “order two.” It is also referred at as a “second-order ordinary differential equation.” More generally, the equation n F x, y, y ', y '',..., y 0 (1.17) is called an “ nth - order’’ ordinary differential equation. Equation (1.17) represents a relation between the n 2 variables x, y, y,..., y solved for y n n which under suitable conditions can be in terms of the other variables: y n f x, y, y,..., y n1 (1.18) For the purpose of this book we shall assume that this is always possible. Otherwise, an equation of the form of equation (1.17) may actually represent more than one equation of the form of equation (1.18). For example, the equation x2 y 3 y 2x 0 actually represents the two different 2 equations, y 3 9 8 x3 2 x2 or y 3 9 8 x3 2 x2 Note: Eq.(1.17) is called the general form. Note also that Eq. (1.17) is an n th order differential equation because: n d y n y n ……………………. dx 14 (1.19) Example 1.2 1. 2. 3. di Ri E dt yy x L (order 1) (order 2) 2 z 2 z 3 0 x2 y 2 (order 2) (c) Degree Definition 1.2 The degree of the differential equation is the power (exponent) or the index that its highestordered derivative is raised, if the equation is rationalized or cleared of fractions with regard to the dependent variable and its derivatives involve in it. From equation (1.11), squaring both sides results in: 3 2 d 3 y dy 3 x 3 e , degree dx dx 2 2 5 d3y d2 y y e x is an ordinary differential equation, The differential equation 3 2 2 dx dx x 1 of order three and degree two. Oral Exercise 2 State the order and degree of equations 1.1, 1.2, 1.3, 1.4, 1.5, 1.6, 1.7, 1.8, 1.9, 1.10, 1.11, 1.12, 1.13, 1.14, 1.15 and 1.16 (d) Linear Differential Equation A differential equation is linear if it can be put in the form: an x y n an1 x y n1 ... a2 x y '' a1 x y ' a0 x y f x (1.20) where an is not identically zero and also the subscripted (or indexed) a ’s are functions of independent variable ( x ) only. The conditions for a linear differential equation are as follows: (i) The dependent variable and all its derivatives occur only in the first degree (or to the first power). (ii) No product of the dependent variable, say y , and/or any of its (iii) derivatives present. No transcendental function (trigonometric, logarithmic or exponential) of the dependent variable and/or its derivatives occurs. 15 Example 1.3 (linear) 1. dy y x2 dx It is linear, it does not matter that the independent variable x is raised to the power dependent and derivative are not. 2. 2 , the 3x2 y 2ln x y e x y 3x cos x This is a second order linear ordinary differential equation Example 1.4 (Non - linear) 1. 4 yy x3 y cos y e2 x This is not a linear differential equation because of the 4yy and the cos y terms. Other examples are: (i) dy y2 0 dx (ii) dy dx 3 y 0 2 3 d 3 y d 2 y dy (iii) 2 e x . 3 dx dx dx Remark: A linear differential equation is always in the first degree of the dependent variable (variables) and the derivatives. Oral Exercise 3 For each of the following, state whether the equations are ordinary or partial, linear or nonlinear, and give its order and degree. 1. 3. d2y k 2x 0 2 dx 2x 2 y 2 dx 2 xydy 0 16 2. 2 2w 2 w a t 2 x2 4. y ' p x y q x dy xy 2 0 dx d4y x dx4 5. y '' 3 y ' 2 y 0 6. 7. 2v 2v 2v 0 x2 y 2 z 2 8. 9. d2y sin x y sin x dx2 10. x2 y yy 0 12. x y '' y ' y 0 14. dy 1 xy y 2 dx 16. d 2 g sin 0 dt 2 l 11. x y dx 3x2 1 dy 0 2 13. 4 d 3w dw 3 2 xw 0 dx dx 15. y '' 2 y ' 8 y x2 cos x 3 4 1-1.3 Notation 4 The expressions y, y, y, y ,..., y n are often used to represent, respectively, the first, second, third, fourth, ..., n th derivatives of y with respect to the independent variable under d2y consideration. Thus, y represents if the independent variable is x , but represents dx 2 d2y 4 n if the independent variable is p . Observe that parentheses are used in y and y to 2 dp y 4 and y n respectively. In mechanics, if the independent variable is time, usually denoted by t , primes are often replaced by dots. Thus, y, y , and y distinguish it from the n th power, dy d 2 y d3y represent , , and , respectively dt dt 2 dt 3 Session 2-1 Solution of a Differential Equation Unlike algebra, in which we seek the unknown numbers that satisfy an equation such as x3 7 x2 11x 41 0 . In solving a differential equation we are challenged to find the unknown functions, say y f x , for which an identity such as f x 2xf x 0 – in Leibniz notation or dy 2 xy 0 holds on some interval numbers. dx 17 Ordinarily, we will want to find all solutions of the differential equation if possible. The solution of differential equations plays an important role in the study of the motions of heavenly bodies such as planets, moons and artificial satellites. Two questions that we will be asking repeatedly of a differential equation in this course are: 1. Is there a solution to the differential equation? 2. Is the solution given unique? 2-1.1 Definition 1.3 A solution of a differential equation is any function, say , satisfying the given differential equation on a specified interval, I. A solution may be defined on the whole real line , or on only a part of the line often a, b . The n derivatives of the function must exist on the interval, say a x b n n 1 x for every x in a x b . such that x f x, x , x ,..., an interval Thus if interval is a solution of some first order differential equation, say dy F x, y, 0 on an dx d I (real), then it implies F x, , 0 . dx Thus to test whether a given function solves a particular differential equation, we substitute the function and its derivatives into the differential equation. If the equation reduces to identity 0 , then the function solves the equation otherwise it does not. It is also important to note that since solutions are often accompanied by intervals and these intervals can impart some important information about the solution. Example 1.5 Show that for any c1 cos x c2 sin x values of the arbitrary constant is a solution of the differential equation c1 and c2 the function d2y y0 dx2 Solution: We will need a second derivative of the solution function, . If is a solution to d2y d 2 F y, 2 0 , then F , 2 0 . If proves otherwise then it is not a solution to dx dx d2y d 2 F y, 2 0 . We differentiate twice to obtain 2 . dx dx 18 d d 2 c1 sin x c2 cos x 2 c1 sin x c2 cos x dx dx d2y d 2 2 y 0 2 0 dx dx d 2 2 c1 cos x c2 sin x c1 cos x c2 sin x 0 dx This implies we have an identity. Hence is a solution to the given ODE and that the d2y formula: c1 cos x c2 sin x gives all possible solution of the equation y 0. dx2 Since sin x and cos x are continuous in the entire real line, the solution is defined in the entire real line , for any arbitrary constant c1 and c2 . Example 1.6 1. 1 2 is a solution of y 2 xy 0 on 1,1 but not on any x 1 larger interval containing I . Show that y 2 Solution: We will need a first derivative of the solution function. y 1 2 x 1,1 . and y 2 are well-defined functions on 2 x 1 x 1 2 Imitating the LHS of the differential equation y 2 xy 2 0 , we have: 2 1 1 y Thus, is a solution of 1,1 . y 2 xy 2 x 0 2 2 x2 1 2 x 1 x 1 2x 2 Note, however, that 1 is not defined at x 1 and therefore could not be a solution on x 1 2 any interval containing either of these two points. 2. Show that on I y ln x is a solution of xy y 0 on I 0, but is not a solution , . Solution: We will need the first and second derivative of the solution function. y ln x , y 1 1 and y 2 are well-defined functions on 0, . x x 19 Imitating the LHS of xy y 0 ,we have: y ln x is a solution on 0, . Thus, Note that 1 1 x 2 0 x x y ln x could not be a solution on , , since the logarithm is undefined for negative numbers containing. 3. Prove that y e x d2y sin x is a solution of 2 y 2e x dx Solution: We will need a second derivative of solution function to do this. y e x sin x , y e x cos x and y e x sin x d2y x y 2 e Imitating the LHS of differential equation dx 2 e x sin x e x sin x 2e x 4. Show that the two functions y c x of the equation x y 2 yc x 1 1 2 2 , y x c x 2 1 2 2 2 x c x x c 2 x 2 5. Show that y x 32 y c2 x2 1 2 are both solutions dy 0, c x c. dx Solution: We will need a first derivative to do this. 2 2 2 and is a solution of 1 2 2 12 0 4x2 y 12xy 3 y 0 for x 0 . Solution: We will need the first and second derivative of the solution function to do this. 3 5 15 7 3 y x 2 , y x 2 and y x 2 are well – defined functions on x 0 or 0, . 2 4 2 Imitating the LHS of 4x y 12xy 3 y 0 , we have: 20 15 7 3 5 4x2 x 2 12x x 2 3x 0 4 2 3 Thus, y x 2 is a solution of x 0 . Note, however, that 3 x2 1 x3 could not be a solution on ,0 , since zero and any negative real number plug into it would give an undefined number and complex number respectively, which is not what we are looking for. Exercise 1.1 In each of assignment 1 – 12, verify by substitution that each given function is a solution of the given differential equation and also state the interval in which the solution exists. Solution(s) Differential Equation 2. y 2e x xe x y 1 y 2 y y 0 y 2 y y x 3. y c 2 cx , c constant 4 x y 2 xy y 0 4. y e x e x y' y 2e x 6. y1 e 2 x , y2 xe2 x y1 x , y2 0 y 4 y 4 y 0 y xy y 0 7. y x x dy 2 y 4 x4 dx xy3 8. y 1. 5. 9. 10. 8 4 2 1 4 1 1 x2 y'2 xy 2 0 1 ln x x y1 x cosln x , y2 x sinln x y1 x ln x , y2 21 x 2 y xy y ln x x 2 y xy 2 y 0 Session 3-1 General Solution 3-1.1 Definition 1.4 A formula that gives all solutions of a differential equation is called the general solution of the equation. A general solution to an n th order DE generally involves n independent arbitrary constants, each admitting a range of real values. NOTE: From the problems of Exercise 1.1, it is clear that the number of distinct solutions depend on the order of the differential equation, for which they are equal. It is worth mentioning that, the differential equation dy 4 describes a straight line with a dx constant gradient of 4 . There are however infinitely constant many straight lines that take the equation y 4x c , where c is an arbitrary constant. Since c R , y is a constantly many hence the differential equation dy 4 has many solutions, because c can assume dx any real number and the differential equation would be satisfied, hence the name arbitrary constant. The general solution is known as family of solutions. y 4x c 3-1.2 Determining a Differential Equation from a General Solution: The Elimination of Arbitrary Constant To determine a differential equation from a general solution, we need to eliminate the arbitrary constant. Methods for the elimination of arbitrary constants vary with the way in which the constants enter the relation. A method which is efficient for one problem may be poor for another. One fact persists throughout. Since each differentiation yields a new relation, number of derivatives that need be used is the same as the number of arbitrary constants to be eliminated. We shall in each case determine the differential equation that is: 22 (a) (b) (c) Of order equal to the number of arbitrary constants in the given relation Consistent with that relation; Free from arbitrary constants. Example 1.7 Find the differential equation whose general solution is y c1e2 x c2e3x Eliminating the arbitrary constants c1 and c2 from the relation: y c1e2 x c2e3 x ______________ (I) Since two constants are to be eliminated, obtain the two derivatives: y ' 2c1e2 x 3c2e3 x __________ (II) y '' 4c1e2 x 9c2e3 x ___________ (III) The elimination of c1 from (II) and (III) yields: y '' 2 y ' 15c2e 3x The elimination of c1 from equations (I) and (II) yields: y ' 2 y 5c2e 3x Hence y '' 2 y ' 3 y ' 2 y Or y '' y ' 6 y 0 . Alternatively, (another) method for obtaining the differential equation in this example proceeds as follows. We know from a theorem in algebra (MATH 152) that three equations (I),(II) and (III) considered as equations in the two unknowns c1 and c2 can have solutions only if: e2 x y e3x y ' 2e2 x 3e3x 0 ______________ (IV) 4e2 x 9e3x y '' Since e 2x factors e and e 2x 3x and e cannot be zero for any x , equation (IV) may be rewritten, with the 3x removed, as: y 1 1 y ' 2 3 0 y '' 4 9 From which the differential equation: y '' y ' 6 y 0 follows immediately. Example 1.8 23 1. Find the differential equation whose general solution is y c sin x , where c is an arbitrary constant. Solution: y c sin x _______________ (I) There is one arbitrary constant, we then find the first derivative: dy c cos x ______________ (II) dx From (I) and (II): c y y sin x cos x y sin x y cos x y sin x y cos x 0 y Alternatively, from (I) let: c sin x y sin x y cos x Differentiating w.r.t. x : 0 sin2 x y sin x y cos x 0 2. Determine the differential equation whose general solution is y c1e x c2e x 2 x . Solution: There are two arbitrary constants, so we get equation of order 2. Hence we differentiate twice. y c1e x c2e x 2 x ……………… (I) dy c1e x c2ex 2 ………………. (II) dx d2y 2 c1e x c2e x ……………… (III) dx dy Add (I) and (II): y 2c1e x 2 x 2 ………. (IV) dx d 2 y dy 2c1e x 2 ………… (V) Add (II) and (III): 2 dx dx Let 24 The two equations above have the same term on the RHS, hence we equate the LHS, and we d 2 y dy dy get: y 2x dx2 dx dx d2y This simplifies as: y 2x 0 2 dx 3. Find the differential equation whose general solution is y c0 c1x c2 x 2 c3 x3 . Solution: The equation has four arbitrary constants c0 , c1, c2 , c3 ; hence we need to differentiate four times to get a differential equation of the fourth order. y c0 c1x c2 x 2 c3 x3 Let d2y dy d3y c1 2c2 x 3c3 x2 2 c 6 c x 6c3 2 3 dx dx2 dx3 4. d4y dx4 0 , which is the differential equation that we need. Eliminate the constant c from the equation Solution: Direct differentiation of the relation yields: 2 x c 2 y2 c2 x c 2 yy 0 from which c x yy . Therefore, using the original equation, we find that: yy2 y2 x yy2 , or x 2 y 2 x 2 2 xyy , which may be written in the form: y 2 dx 2 xydy 0 Alternatively: The equation x c 2 y2 c2 may be put in the form: x2 y 2 2cx 0 or x2 y 2 2c then differentiation of both sides leads to: x x 2xdx 2 ydy x2 y 2 dx x 2 0 or x2 y 2 dx 2 xydy 0 25 5. Eliminate B and from the relation x B cos t in which is a parameter (not to be eliminated) Solution: First we obtain two derivatives of x with respect to t : dx d 2x B sin t and 2 2 B cos t dt dt 2 d x 2 2 x , since x B cos t dt d 2x 2 2x 0 dt Eliminate c from the equation cxy c x 4 0 2 6. Solution: At once we get: c y xy c2 0 Since c 0 , c y xy and substitution into the original equation leads us to the result: 2 x3 y x2 yy 4 0 Remark: The general solution of an n th order ordinary differential equation could be expected to have n arbitrary constants. Exercise 1.2 Assuming that a, b, c, c1, c2 , A, , B are arbitrary constants, show that each function or equation on the left satisfies the differential equation written opposite it. Remember that we differentiate and substitute. Solution Differential Equation 1. x y c yy x 0 2. y cx c4 y xy y 3. y c1x c2e x 4. x3 3 x 2 y c 5. y sin x xy 2 c 2 2 4 x 1 y xy y 0 x 2 y dx xdy 0 y cos x y dx sin x 2xy dy 0 26 6. x 2 y 1 cx 7. cy 2 x2 y x y 1 dx x dy 0 2 xydx y 2 x dy 0 2 3 2 d 2x 2x 0 2 dt d 2x 2x 0 2 dt 8. x Asin t , : parameter 9. x c1 cost c2 sin t , : parameter 10. y cx c 2 1 11. y cx 12. y 2 4ax 2xdy ydx 0 13. y c1e px c2e px ; p R y p2 y 0 y xy y 1 2 y xy c 1 y 2u 2u 0 x 2 y 2 y x 14. u tan 15. u log x y 16. y xc cos x 17. y e x 1 x y 2 y y 0 18. y ax2 bx c y 0 19. y c1 c2e2 x y 2 y 0 20. y 4 c1e2 x y 2 y 8 21. y c1e x c2e3 x y 4 y 3 y 0 22. y c1e x c2 xe x y 2 y 2 y 0 23. y x2 c1e x c2e2 x y y 2 y 2 1 x x 2 24. y c1e2 x cos3x c2e2 x sin3x y 4 y 13 y 0 25. y c1x 2 c2e2 x 2 2 1 2 uxx u yy 0 xy y x 2 sin x x 1 x y 2 x2 1 y 2 2 x 1 y 0 27 Session 4-1 Initial-value and Boundary-value Problems 4-1.1 Definition 1.5: Subsidiary Conditions Subsidiary Condition(s) is/are condition(s), or set of conditions, on the differential equation that will allow us to determine which solution that we are after. 4-1.2 Definition 1.6: Initial–Value Problem Initial-Value Problem is the differential equation along with subsidiary conditions on the unknown function and its derivatives, all given at the same value of the independent variable. The subsidiary conditions are initial conditions. 4-1.3 Definition 1.7: Boundary–Value Problem Boundary-Value Problem is the differential equation along with subsidiary conditions on the unknown function and its derivatives, which are given at more than one value of the independent variable. The subsidiary conditions are boundary conditions. Theorem 1.1 An initial value problem (IVP) for an n th order DE includes a specification of the solution’s value and n 1 derivatives at some point or a set of n algebraic conditions at a common point: yx0 y0 , dy x0 y1, dx ..., d n 1 y x0 yn 1 dxn 1 Generally in applications, an IVP has a unique solution on some interval containing the initial value point. Example 1.9 The problem y 2 y e ; y 1, y 2 is an initial – value problem, because the x two subsidiary conditions are both given at x . The problem y 2 y e ; y 0 1, y 1 1 is a boundary – value problem, because the x two subsidiary conditions are given at the different values x 0 and x 1 . If an initial condition is given along with the differential equation, that is, a constraint of the form y y0 when x x0 , then this information can be used to determine the particular value of c . In this way, one particular solution (actual solution) can be selected from the family of solutions – the one that satisfies both the differential equation and the initial conditions. 28 The geometric interpretation of the general solution of a first order differential equation dy F x, y, 0 is an infinitely many family curves; one for each value of real number. dx These curves are called integral curves, sometimes it is important to pick one particular member of the family of integral curves. This is done by identifying a particular point x0 , y0 through which the graph of the solution is required to pass. This requirement is usually written as y x0 y0 . This is referred to as an initial value condition and the problem of the form: dy F x, y, 0; y x0 y0 …………………. (1.21) dx is called initial value problem. Note: If initial condition(s) is used to solve a differential equation, the solution is then called particular solution (actual solution). Example 1.10 1. y c1e x c2e3 x is a general solution of the differential equation: y 2 y 3 y 0 Determine the particular solution of the initial conditions y 3 when x 0 and y 4 when x 0 . Solution: Since the differential equation is second order, we have a specification of the solution y(0) 3 and one derivative at a point: y 0 4 , that is 2 conditions at a common point: x 0 y c1e x c2e3 x when x 0 , y 3 let value: 3 c1e 0 c2e3 0 c1 c2 ______(I) y c1e x 3c2e3x when x 0 , y 4 , 4 c1e 0 3c2e 0 4 c1 3c2 _______________ Solving equation (I) and (II) we have, c1 29 (II) 5 7 and c2 4 4 algebraic the particular solution is given by: 5 7 1 y e x e3x or y 5e x 7e3x 4 4 4 2. Find a solution to the boundary – value problem y 4 y 0; y /8 0, y / 6 1, if the general solution to the differential equation is y c1 sin 2 x c2 cos2 x . Solution: For x 8 , y c1 sin2 8 c2 cos2 8 c1 sin 4 c2 cos 4 2 2 c2 2 2 To satisfy the condition y /8 0 , we have: y c1 2 2 c2 0 ………………….(I); 2 2 For x 6 , y c1 sin2 6 c2 cos2 6 c1 sin 3 c2 cos 3 c1 3 1 c2 2 2 To satisfy the condition y / 6 1 , we have: y c1 c1 3 1 c2 1 ……………………..(II) 2 2 Solving (I) and (II) simultaneously, we get: c1 2 2 and c2 . 3 1 3 1 Hence a particular solution is obtained: y 2 sin 2 x cos2 x 3 1 as the solution of the boundary – value problem. 3. Find a solution to the boundary – value problem y 4 y 0; y 0 1, y / 2 2 , if the general solution to the differential equation is known to be y c1 sin 2 x c2 cos2 x . 30 Solution: For x 0 , y c1 sin0 c2 cos0 c2 To satisfy the condition y 0 1 , we have: c2 1 …………………........................(I). For x / 2 , y c1 sin 2 c2 cos 2 c2 y / 2 2 , we have: c2 2 c2 2 ………………….(II). To satisfy the condition Thus, to satisfy both boundary conditions simultaneously, we must require c2 to equal both 1 and 2 , which is impossible. Therefore, there is no solution to this problem. Exercise 1.3 1. Verify that the following are solution of the given differential equations. a) for y 2 sin 2x y 4 y 0 d) y e3 x 1 y 1 x y e x cos x e) y 2 x x ln x b) c) 2. for y 3 y 0 for y y2 0 for y y 2 y 0 for 4x2 y y 0 Determine c1 and c2 that y c1 sin 2 x c2 cos2 x 1 will satisfy the conditions y /8 0 and y /8 2 . 3. Determine c1 and c2 that y 0 0 and y 0 1. 4. y c1e2 x c2e x 2sin x will satisfy the conditions In Problems 4 a) through e), find values c1 and c2 so that the given functions will satisfy the given conditions. Determine whether the given conditions are initial conditions or boundary conditions: a) y c1ex c2e x 4sin x ; y 0 1, y 0 1 b) y c1x c2 x2 1; y 1 1, y 1 2 c) y c1 sin x c2 cos x y 0 1, y 2 1 d) y c1 sin x c2 cos x y 4 0, y 6 1 e) y c1e x c2 xe x x2e x y 1 1, y 1 1 31 Example 1.12 Show that the integral curves of the differential equation: y x dx y 3 3 x dy 0 are given by the family y 4 xy x c . 4 4 Solution: Apply implicit differentiation to the proposed family of integral curves to find y ; y 4 4 xy x4 c 4 y3 y ' 4xy ' 4 y 4x3 0 4 y 3 y 4xy 4 y 4x3 0 y y 3 x y x3 0 dy 3 y x y x3 0 dx dy y x3 3 dx y x This last equation can then be written in terms of the differentials dx and dy y3 xdy y x3 dx y x3 dx y 3 x dy 0 Notice that the last equation can be put in the form M x, y dx N x, y dy 0 , which is simply an alternate form of the equation: dy M x, y dx N x, y where M x, y y x3 and N x, y y3 x . Summary - Definition of Differential Equation: A differential equation is an equation which involves one or more derivatives, or differentials of an unknown function. - Classification of Differential Equation: o Types o Order o Degree o Linearity 32 - A differential equation is linear if it can be put in the form: an x y n an1 x y n1 ... a2 x y '' a1 x y ' a0 x y f x where an x is not identically zero and also the subscripted (or indexed) a ’s are functions of independent variable ( x ) only. The conditions for a linear differential equation are as follows: o The dependent variable and all its derivatives occur only in the first degree (or to the first power). o No product of the dependent variable, say y , and/or any of its derivatives present. o No transcendental function (trigonometric, logarithmic or exponential) of the dependent variable and/or its derivatives occurs. 33 UNIT 2 ORDINARY DIFFERENTIAL EQUATION OF FIRST ORDER Introduction In this unit we begin our program of solving differential equations. The first – order differential equations involve only the first derivative of the unknown function. Our usual notation for the unknown function will be y x . Learning Objectives After going through this unit, you would be able to: • put first order ordinary differential equation in general, standard and differential forms • find the difference between all first all the differential equation. • describe the methods of solving most first order differential equations. Session 1-2 First Order Differential Equations First order differential equations are often used to model the behaviour of engineering systems. For example, in a radioactive decay, mass is converted to energy by radiation. It has been observed that the rate of change of the mass is proportional to mass itself. That is if N t is the mass at time t then: dN t dt N t dN t dt kN t for some k that depends on the element. Such an equation is called a Model in mathematics. 1-2.1 Definition 2.1 A first order differential equation is an equation containing only the first differential. In its general form the 1st order differential equation is given by: F x, y, F x, y, y 0 dy 0 or dx The above equation has x called the independent variable, the unknown function y which depends on x and the first derivative of y (i.e. y ) with respect to x . 34 Examples of first-order differential equations dy 5x 0 dx dy 4 xy x 3e x 2. dx dy y0 3. x dx dy 4 xy 4. 3 dx 1. Standard form for a first order – differential equation in the unknown function y is: y f x, y ……………………….. (2.1) where the derivative y appears only on the left side of (2.1). Many, but not all, first – order differential equations can be written in standard form by algebraically solving for y and then setting f x, y equal to the right side of the resulting equation. Example 2.1 1. Write the differential equation xy y 2 0 , which is in the general form, in standard form. Solution: Solving for y , we obtain: 2. y y 2 / x as standard form with f x, y y 2 / x Write the differential equation ex y e2 x y sin x general form and standard form. Solution: Shifting everything to the left of equation we have: and solving for y , we obtain: ex y e2 x y sin x 0 as general form y ex y e x sin x as standard form, with f x, y e x y e x sin x . 3. Write the differential equation y y sin y / x in standard form. 5 Solution: This equation cannot be solved algebraically for y , and cannot be written in standard form. 35 Differential form: The right side of (2.1) can always be written as a quotient of two other functions M x, y and N x, y . Then (2.1) becomes: y M x, y / N x, y , or dy M x, y which is equivalent to the differential form: dx N x, y M x, y dx N x, y dy 0 ……………………….. (2.2) where we call the functions M x, y and N x, y the coefficients of dx and dy respectively. NOTE: We will put most (if not all) first order ordinary differential equation in the form of equation 2.2 before solving. Example 2.2 Write the differential equation y yy 1 x in the differential form. Solution: Solving for y , we have: y with f x, y 2 y y x y x y ……….......................... y2 x y , which is the standard form y2 (I) There are infinitely many different differential forms associated with (I), which are as follows: b) Take M x, y x y, N x, y y . 2 Then dy M x, y x y x y 2 the equivalent differential form is: 2 dx N x, y y y x y dx y2 dy 0 c) y2 . x y Take M x, y 1, N x, y Then dy M x, y 1 x y 2 2 the equivalent differential form is: dx N x, y y / x y y y2 1 dx dy 0 x y 36 x y y2 d) Take M x, y . , N x, y 2 2 dy M x, y x y / 2 x y 2 the equivalent differential form is: Then dx N x, y y 2 / 2 y y2 x y 2 dx 2 dy 0 x y y2 , N x, y 2 . e) Take M x, y x2 x 2 dy M x, y x y / x x y Then the equivalent differential form is: dx N x, y y 2 / x2 y2 y2 x y x2 dx x2 dy 0 Exercise 2.1 1. 2. dy y in differential form. dx x 2 Write the differential equation xy 3 dx 2x y 1 dy 0 in standard Write the differential equation form. xy y 2 0 in differential form. 3. Write the differential equation 4. Write the differential equation y 5 y 6 x y y 2 in standard 2 form. y y x in differential form. 5. Write the differential equation e 6. Write the differential equation y y y sin x in differential form. 3 37 2 1-2.2 General and Particular Solutions of a First Order Ordinary Differential Equation In this section, we examine methods of solving a first order differential equation. The methods or differential equation kinds to consider are as follows; (I) Method of separation of variables (II) Reduction to Separable Form Homogeneous equation Transformations (coefficients linear in the two variables) (III) Exact Differential Equations (IV) Method of Integrating – factor used for the following: a) Non – exact differential equation b) Linear ordinary differential equation of first order and those reducible to that form: (i) Bernoulli’s Equation (ii) Riccati’s Equation Theorem 2.1: An Existence Theorem for Equations of Order One Consider the equation of order one dy f x, y ………………………………… (2.1) dx Let T denote the rectangular region defined by: x x0 a and y y0 b , where a, bR , a region with the point x0 , y0 at its centre. Suppose that f and f / y are continuous functions of x and y in T . Under the conditions imposed on f x, y above, there exists an interval about x0 , x x0 h , and a function y x which has the properties: (a) y y x is a solution of equation (2.1) on the interval x x0 h ; (b) On the interval (c) At x x0 , x x0 h y y x0 y0 ; , y x satisfies the inequality y x y0 b ; y x is unique on the interval x x0 h in the sense that it is the only (d) function that has all of the properties (a) , (b), and (c). The interval x x0 h may or may not need to be smaller than the interval x x0 a over which conditions were imposed upon In other sense, the theorem states that x0 , y0 then the differential equation f x, y . f x, y is sufficiently well behaved near the point 38 dy f x, y …………………………………….(2.1) dx has a solution that passes through the point x0 , y0 and that the solution is unique near x0 , y0 . Session 2-2 Method of Separation of Variables Some differential equations have a special form that allows us to separate them. Such equations can be solved by using the method in this section. A separable differential equation is a first-order ordinary equation that is algebraically reducible to a standard differential form in which each of the non-zero terms contains exactly one variable. Solutions to this kind of equation are usually quite straightforward. 2-2.1 Definition 2.2 Separable Differential Equation A first order differential equation is said to be separable if it can be written in the form: dy f x, y G x H y or equivalent form: dx M x, y dx N x, y dy 0 ------------- (2.2) If (2.2) is separable, then: M x, y m x p y and N x, y n x q y 2.3 Then (2.2) is called separable differential equation; m x p y n x q y We then divide throughout by dy 0 ...……………… (2.4) dx p y n x , we get: m x p y dy 0 n x q y dx m x p y dx dy 0 …………………………….. (2.5), n x q y now it is separated in each variable x and y (hence the method’s name) 39 Integrating both sides: m x p y dx dy c ………………………. n x q y (2.6) where c is arbitrary constant. (2.6) is then the general solution of the differential equation. The evaluated term on the left-hand side of (2.6) is a function F whose total differential is the left-hand side of (2.5) ALGORITHM 1 Separation of Variables: To solve the initial-value problem y f x, y ; y x0 y0 ……………………………… (2.7) (a) Separate the variables, so that all of the x -dependence is on one side of the equation, and all of the y -dependence is on the other side of the equation. (b) Integrate both sides of the equation. (c) Substitute the initial condition to solve for the constant of integration. Example 2.3 1. Solve the following differential equations by separation of variables: (a) (b) dy x y 2 1 dx dy x2 dx y 1 x3 x 1 Solution: (a) We change to differential form, separate the variables, and integrate: x 1 dy x y 2 1 dx x 1 dx 2 dy 0 x 1 y 1 x 1 Assuming x 1 , integrating both sides: dx 2 dy c we get: x 1 y 1 x log x 1 tan1 y c , as our general solution. ln Note: is the same as log 40 dy x2 x2 dx and assuming that x 1 (b) The ODE becomes: ydy dx y 1 x3 1 x3 , then: y2 1 x2 ln 1 x3 c dx 3 2 3 1 x 2 2 y ln 1 x3 2c ln 1 x3 c1 3 3 ydy Since c is arbitrary constant, any number multiplying or dividing it will not have any effect, it will still be arbitrary constant. We replace 2c with c1 to bring a distinction but no effect in general, so one can write the general solution as: 2. Solve the equation (IVP) y 2 ln 1 x3 c 3 1 x2 x1 2 yy 0 y1 0 Solution: 1 x2 x1 2 yy 0 1 x2 dy 1 2 y 0 Divide throughout by x x dx Assuming that x 0 , the general solution is: 2 1 x dx x 1 2 ydy c x2 ln x y y2 c 2 To solve for the arbitrary c , we substitute x 1, y 0 ; 12 1 ln1 0 02 c c 2 2 x2 1 ln x y y 2 : as particular(actual)solution. 2 2 dy x2 3. Find a particular solution of , y 2 1 dx 1 y 2 41 Solution: The variables are separable, since the equation can be written as: Integrating both sides: x dx 1 y dy 0 , we get: 2 x2dx 1 y2 dy 0 2 x3 y3 y c x3 3 y y3 3c c 3 3 Note the replacement of 3c with c Solving for arbitrary constant, c : 2 3 1 c c 4 3 Then particular solution is given by: x 3 y y 4 . 3 3 Exercise 2.2 Solve the differential equations Ans. 2 x y2 c c c 18 9 4 1. 9 yy 4x 0 2. ln x 3. x cos xdx 1 6 y2 dy 0; y 0 x sin x cos x 1 y6 y mydx nxdy xy y 0; y 1 1 xm cy n 4. 5. 8. 9. 10. 11. 1 2 ln x 2 ln y c y 1/ x cosh 1 y sinh 1 x c 1 x2 dy y 2 1dx 0 6. 7. dx x dy y dy ex y e y ex c dx x2 1 x2 y y xe dy dx 0 y 1 e ln x c y 2 dV V PV C dP P dr r r0 exp 2t 2 4rt ; when t 0, r r0 dt xy 2dx e x dy 0 ; when x , y 12 y e x / 2e x x 1 12. d g ; when x x0 , 0 dx 42 2 02 2g x x0 2-2.2 Reduction to Separable Form Certain differential equations are not separable but can be made separable by the introduction of a new unknown function. We illustrate the idea of this method by some typical cases, i.e. Homogeneous forms and Non-Homogeneous forms, where the coefficients are linear in the two variables, transform to Homogeneous form. a) Homogeneous Functions Polynomials in which all terms are of the same degree, such as: x2 3xy 4 y 2 , x3 y 3 , x 4 y 7 y5. are called homogeneous polynomials. We wish now to extend that concept of homogeneity so it will apply to functions other than polynomials. A formal definition of homogeneity is: A function of two variables, f x, y , is said to be homogeneous of degree n if there is constant n such that: f x, y n f x, y …………………………………… (2.8) for all , x and y for which both sides are defined. Example 2.4 1. f x, y x 4 x3 y f x, y x 4 x 3 y 4 x 4 4 x3 y 4 x 4 x3 y f x, y 4 f x, y Thus the equation of is homogeneous with degree 4. 2. f x, y e y x f x, y e e y y x y tan x x y tan x y tan x f x, y 0 f x, y This function is homogeneous of degree zero. 43 3. f x, y x 2 sin x cos y f x, y x 2 sin x cos y 2 x 2 sin x cos y 1 f x, y 2 x 2 2 sin x cos y f x, y 2 f x, y This function is not homogenous. 4. 2x y f x, y f x, y x2 y2 2x y x2 y 2 2x y 2 x 22 y 2 2x y 4 x 2 y 2 2x y f x, y 3 2 2 x y f x, y 3 f x, y This function is homogeneous of degree 3 . Equations with homogeneous coefficients A differential equation of the form (differential form): M x, y dx N x, y dy 0 …………………….. (2.9) is homogeneous if the functions M x, y and N x, y are both homogeneous functions of the same degree. Or if the equation in standard form: y f x, y ………………………….. (2.10) depends only on the ratio y x or , then it is said to be homogenous. x y To solve homogeneous equations, turn them into separable ones using the substitution: y x …………………………. (2.11), where is a function of x ; also: dy d x ……………….. dx dx or dy xd dx 44 (2.12) Example 2.5 1. Solve the differential equation x 2 y 2 dx 2xydy 0 Solution: Notice that the function M x, y x 2 y 2 and N x, y 2 xy are both homogeneous of degree 2; therefore, the differential equation is homogeneous. The substitution y x implies that dy xd dx , and given equation is transformed into: x2 xv 2 dx 2 x x xd vdx 0 After little algebra, this becomes the separable equation: 1 2 dx d x 1 2 Integrating both sides gives: log x log 1 2 log c , in which we wrote the arbitrary constant of integration as log c because it makes the next step neater: log x log c 1 2 Finally, to write the solution in terms of the original variables, x and y , we replace by y x: x 2. c 1 y / x 2 x 2 y 2 cx Solve the differential equation xdy ydx x 2 y 2 dx 0 Solution: The differential equation is clearly homogeneous degree 1. Using the transformation y x , dy dx xd and dividing by x , we have: dx xd dx 1 2 dx 0 d xd 1 2 dx 0 Integrating both sides: d 1 2 1 2 dx 0 x dx x 1 We get: sin ln x ln c And returning to the original variables, using 45 y/x We get: 3. sin 1 x ln xc xc exp sin 1 y / x y Solve the differential equation xy y x y . 3 Solution: 3 y y 2 y Express the differential equation in form of : y x 1 x x x y Substitute and y x by equation (2.11) and (2.12) on page 34, we get: x dv x x2 1 x x 2 1 dx 1 x2 dv xdx log 1 c 1 2 y Returning back to the original variables: , we get: x x2 y 1 exp c x 2 x2 y x 1 exp c 2 Exercise 2.3 Solve the following differential equations (IVP) 1. xy dy x 2 y 2 ; y 1 2 ; dx Ans. y 2 x2 ln x2 4x2 or x 2 ln x 2 4 x 2 Note the negative square root is taken to be consistent with the initial condition xy y x 2. y x ln xc 3. xy y x 2 sec x ; y1 4. xy y 3x 4 cos2 y / x; y1 0 5. xyy 2 y 2 4x 2 ; y2 4 y 46 A Transformation (Equations in which M x, y and N x, y are linear but not homogeneous) Another type of equation that can be reduced to a more basic type by means of a suitable transformation is an equation of the form: a1x b1 y c1 dx a2 x b2 y c2 dy 0 ……….. (2.13) where a1 , b1 , c1 , a2 , b2 , c2 are constants where M x, y a1x b1 y c1 and N x, y a2 x b2 y c2 written as M x, y N x, y y 0 , which are not homogeneous. However the differential equation can be made homogeneous and solved accordingly provided; Case I a2 b2 , then the transformation: x X h and y Y k , where h, k is a1 b1 the solution of the system a1 x b1 y c1 0 and a2 x b2 y c2 0 , that is x h and y k , which reduces (2.13) to a homogeneous equation: If a X b Y d X a X b Y dy 0 …..(2.14) 1 1 2 2 in the variables x and Y . Case II a2 b2 k , then the transformation z a1x b1 y reduces the equation a1 b1 (2.13) to a separable equation in the variables x and z : If dz dy a1 b1 ………………………… dx dx (2.15) Example 2.6 1. Solve the differential equation: x 2 y 1dx 4 x 3 y 5dy 0 ………… (I) Solution: Here a1 1, b1 2, a2 4, b2 3 and so case I transformation is applied. 47 a2 4 a1 and b2 3 a2 b2 , b1 2 a1 b1 Hence, we then solve: x 2 y 1 0, 4x 3y 5 0 simultaneously to obtain: x 3, y 2 . This implies the solution of the system is h 3, k 2 , and so the transformation is: x X 3 and y Y 2 This reduces the (I) to the homogeneous equation: X 2Y dX 4X 3Y dy 0 …………… (I) dy 1 2 d x 3 4 y x y x Let Y X to obtain: d 1 2 3 4 d d x ….. (II) d x 3 4 1 2 3 2 x 3 4 d d x Integrating both side: 1 2 3 2 x x 3 4 d 15 d 1 d d x 4 1 x 1 3 1 4 1 3 5 1 ln 1 3 ln 1 ln x ln c1 4 4 5ln 1 3 ln 1 4ln x 4lnc1 1 4 4 5 x c1 1 3 1 c 1 3 x4 , where c c14 5 These are the solutions of the separable equation (II). y Now, substituting , we obtain the solutions of the homogeneous equation (I) in the x form: x y c x 3y Finally, replacing X by 5 x 3 and y by y 2 , from the original transformation, we obtain the solution of the differential equation in the form: x 3 y 2 cx 3 3 y 25 x y 1 cx 3 y 95 2. If x 2 y 3dx 2 x 4 y 1dy 0 , determine the general solution. 48 Solution: Here a1 1 , b1 2 , a2 2 , and b2 4 a2 b2 2, a1 b1 therefore it follows case (II) So, let z x 2 y , z 1 2 y The transformations give us: z 3dx 2 z 1 dz dx 0 2 7dx 2 z 1dz 0 This is separable. 7dx Now, replacing 2 z 1dz c 7 x z 2z c z by x 2 y , we obtain the solutions as: 7 x x 2 y 2 x 2 y c Exercise 2.4 Solve each differential equation by making a suitable transformation. 1. 2. 3. 4. 5x 2 y 1dx 2 x y 1dy 0 3x y 1dx 6 x 2 y 3dy 0 x 2 y 3dx 2 x y 1dy 0 10x 4 y 12dx x 5 y 3dy 0 Solve the following initial-value problems 5. 6. 7. 8. 6x 4 y 1dx 4x 2 y 2dy 0; y12 3 3x y 5dx x y 2dy 0; y2 2 2 x 3 y 1dx 4 x 6 y 1dy 0; y 2 2 4 x 3 y 1dx x y 1dy 0; y3 4 49 2-2.3 Exact Equation Given a function f x, y , its total differentials, df , is defined as: f f dx dy ---------------------(٠) x y This shows that the family of curves (or general solution) f x, y c satisfies the differential equation df 0 . f f dx dy 0 ----------------------(٠٠) x y So if there exists a function f x, y such that: df M x, y f f and N x, y y x then M x, y dx N x, y dy is called an exact differential, and the equation: M x, y dx N x, y dy 0 ……………………….(2.2) is said to be an exact equation, whose solution is the family f x, y c . How can you tell when a differential equation is exact? Recall that in partial derivatives if f x, y continuous second partial derivatives, then: 2 f 2 f (that is, the order of xy yx differentiation does not matter). So if f M N f 2 f 2 f N , then: M and (which is ) x y x y xy yx Not only is this condition necessary for exactness, it also determines exactness. Example 2.6 1. Solve the differential equation 1 2 xy dx 4 y 3 x 2 dy 0 . Solution: With M x, y 1 2 xy and N x, y 4 y 3 x 2 , we notice that: M N 2 x y x So the given functions M and N pass the test for exactness. All that is left to do is to find the function f x, y such that: f f N. M and x y 50 The way to accomplish this is to integrate M with respect to x , integrate N with respect to y , and then merge the results: M x, y dx 1 2xydx x x y (a function of y alone) 3 2 4 2 N x, y dy 4 y x dy y x y (a function of x alone) These calculations imply that a function f x, y which satisfies both: 2 f f N 4 y 3 x 2 is: f x, y x x 2 y y 4 M 1 2 xy and x y Therefore, the exact equation we were given is satisfied by the family of curves: x x2 y y 4 c Solve the differentiation equation y cos x 2 xe y sin x x 2e y 2 2. Solution: 0 dy dx M x, y y cos x 2 xe , N x, y sin x x e 2 y M y cos x 2 xe , N 2 y cos x 2 xe y y x Therefore, the differentiation equation is exact. We know that f f N x, y . Thus there is an f x, y , such that: M x, y ; x y f x y y cos x 2 xe , and f y Integrating the first of these equations: df sin x x e 2 ….(I) 2 y y cos x 2xe y dx , keeping y constant gives: f x, y y sin x x e y …………….(II) 2 y and hence differentiating the new function w.r.t f y sin x x e 2 y d dy y keeping x constant: sin x x e 2 , 2 y i.e. the differential is equal to the second equation in (I). d 2 d 2dy and 2y the constant of integration can be omitted; we do dy not require the most general one. Thus Substituting for y in (II) gives: f x, y y sin x x e 2 y 2 y 51 Before f x, y c ,constant. Hence 3. c y sin x x 2e y 2 y Solve the differential equation 3x2 2xy x y2 dy 0 dx Solution: Here M x 2 x , N x 1 ; since M y N x , the given equation is not exact. To see that it cannot be solved by the procedure above, let us suppose that there is a function f x, y such that: f x 3x 2 2 xy , f y x y 2 ………….(I) Integrating the first of (I) gives: f ( x, y) x3 x 2 y y ………………(II) where y is an arbitrary function of y only. To try to satisfy the second of (I) we compute f y from (II), obtaining: f y x2 d x y2 dy d x y 2 x 2 ……………(III) dy Since the right – hand side of (III) depends on x as well as y , it is impossible to solve (III) for y . Thus there is no f x, y satisfying both of (I) or So in the next sub – topic we try to find a factor to make the equation exact and have a function satisfying both (I) – the factor is called integrating factor. Session 3-2 Integrating factors 3-2.1 Non- Exact Equations If M x, y dx N x, y dy 0 .................. (2.2) is not exact, sometimes we can turn it into an exact differential equation by multiplying the whole equation by an appropriate factor, called an integrating factor, let’s say x, y , which is a function of x, y . For which we obtain: M x, y dx N x, y dy 0 ---------------- (2.13) and expect it to be exact. Then for (2.13) to be exact then: 52 M N y x Applying product rule we get: M N M N y y x x N M ------------- (2.14) N y x x y M y N x N x M y M Simply put: N x M y M y N x Remark: Finding the integrating factor for a given equation can be very difficult. Some of the important rules and procedures follow. Theorem 2.1 a) d M y Nx y 0 ,then: x , which is a y dx N function of x alone, call this function x . Assume x x e x dx is an integrating factor of (2.2). d M y Nx 0 , then: y Assume y , which is a x dy M function of y alone, call this function y . Then b) Then y e c) If y dy is an integrating factor of (2.2). 1 M N 0 , then: x, y , is the integrating factor of M N x y x y (2.2). d) If M x, y dx N x, y dy 0 can be written in the form yg x, y dx xh x, y dy 0 , where f x, y g x, y , then x, y 1 xM yN 53 The following observations are often helpful in finding integrating factors: (1) If a first-order differential equation contains the combination 1 xdx ydy d x2 y 2 try some function x2 y 2 as a multiplier. 2 (2) If a first-order differential equation contains the combination xdy ydx d xy , try some function xy as a multiplier. (3) If a first-order differential equation contains the combination xdy ydx , try 1 1 or x2 y 2 1 1 or 2 or some function of these xy x y 2 xdy ydx 1 y expressions, as an integrating factor, remember that d tan and xy x y xdy ydx . tan 1 2 x y2 x as a multiplier. If neither of these works, try Sometimes an integrating factor may be found by inspection, after grouping the terms in the equation, and recognizing a certain group as being a part of an exact differential. The table below has some of the forms already listed above: 54 Table 2.1 Group of Terms Integrating factor ydx xdy ydx xdy Exact differential xdy ydx y d 2 x x 1 x2 x ydx xdy d 2 y y xdy ydx y d ln xy x xdy ydx y d arctan x2 y 2 x ydx xdy d ln xy xy 1 y2 ydx xdy ydx xdy ydx xdy 1 xy 1 x y2 2 1 xy 1 ,n 1 xy n ydx xdy ydy xdx ydx xdy 1 d n1 xy n n 1 xy ydy xdx 1 d ln x2 y 2 2 2 x y 2 1 x y2 2 ydy xdx x aydx bxdy a, b constant 1 2 df x, y y 2 n ,n 1 xa1 yb1 ydy xdx x 2 y2 n 1 d 2 n 1 x 2 y 2 n1 xa1 yb1 aydx bxdy d xa yb NOTE: If a non – exact equation has a solution, then an integrating factor is guaranteed to exist, but that does not mean it is easy to find. Example 2.7 1. Solve the differential equation Solution: 3xy y2 x2 xy dy 0. dx M ( x, y) 3xy y 2 , N x, y x 2 xy 55 M y 3x 2 y and N x 2x y . M N , therefore the differential equation is not exact. y x Involve (2.14): M y N x N x M y Clearly, Let be a function of x alone, that is, Applying theorem 2.1a: x M y Nx N x , 3x 2 y 2 x y 1 x x 2 xy x e x dx e x 1 dx Multiplying eln x x x x through the differential equation we obtain: dy 3x 2 y xy2 x3 x 2 y 0 dx which is a new equation in same form i.e.: M x, y N x, y y 0 M x, y 3x 2 y xy 2 , N ( x, y) x3 x 2 y M y 3x 2 2 xy , N x 3x 2 2 xy ,which is exact. f x 3x 2 y xy 2 Let Integrating: f x, y 3x 2 y xy 2 dx x3 y 12 x 2 y 2 y Find the derivative w.r.t y and equating with N x, y we obtain: f y x3 x 2 y y x3 x 2 y y 0 , x 3 y 12 x 2 y 2 c . Finally 2. Solve Solution: y c y 2 y dx xdy 0 . M x, y y 2 y , N x, y x M y 2 y 1 Nx 1 56 Hence the differential equation is not exact, and no integrating factor is immediately apparent. Note, however, that if terms are strategically regrouped, the differential equation can be rewritten as: ydx xdy ydx 0 (I) The group of terms in parenthesis has many integrating factors in table (2.1). Trying each integrating factor separately, we find that the only one that makes the entire equation exact is: x, y 1 . y2 Using this integrating factor we can rewrite (I) as: ydx xdy 1dx 0 y2 (II) Since (II) is exact, it can now be solved. Alternatively, we note from table 2.1 that (II) can be rewritten as x d x / y 1dx 0 , or as d 1dx . y Integrating, we obtain the solution: x x x c or y xc y Exercise 2.5 Solve the following differential equation Ans. 1. 2xy 1 x y 0 x y y c 2. y xy 0 y cx 3. x sin y x cos y 2 y y 0 4. y xy 0 1 2 x x sin y y 2 c 2 xy c 5. 2 2 3x2 y2 2x3 y 4 y3 y 0 6. x 7. ydx 1 x dy 0 8. 2xy y 2dy 0 y 2 2 c x2 9. 1 2 xyy 0 y 2 ln cx 10. x 1 14 1 y x2 3 3 x 2 2 y y2 dx xdy 0 x3 y 2 y 4 c y x tan x c cy x 1 y dx xdy 0; y 1 5 57 3-2.2 Linear Equations and those Reducible to that form A first – order linear equation is defined as a differential of the form: A( x) y Bx y Cx , ……………………….. (2.15) where Ax 0 such that the derivative of the dependent variable exist otherwise it is not a differential equation. Then (2.14) in the standard form is written as y px y qx ……………………………… (2.16) p x To divide by B x C x and qx Ax Ax Ax would violate the fundamental commandment of mathematics, if Ax 0 , which prohibits division by zero. We recognise (2.16) as an inexact differential equation which requires an integrating factor x , that is: d x 0 dy x to obtain the following: xyx pxyx qx 0 ……………… (2.17) M x, y x px y qx and N x, y x Multiplying through (2.16) by M N . y x py q x d x y x dx d p …………. (2.18) dx If (2.17) is exact, then Thus, Separating variables: d pdx ln pdx e pdx …………………….. (2.19) Without loss of operating the constant of integration of (2.19) is unity. Consider (2.16) and multiply through by: x e px dx , then we have the exact equation: y py q y y q by (2.18) that is p d p p dx 58 Integrating both sides: y qdx c y 1 qdx 1c p x dx p x dx p x dx ye qxdx ce …… (2.20) e The solution is in two parts in the R.H.S. The term ce p x dx 1c , called the complementary function is the solution of the reduced equation: y px y 0 obtained by setting qx 0 in (2.16). The differential equation which has qx 0 is called Homogeneous differential equation. The other term on the R.H.S of the solution is dependent on qx , that 1 qdx is called the particular integral and it satisfies (with the complementary function) the differential equation: y px y qx , called the Non – Homogeneous differential equation. Notice that the particular integral contains no constant. ALGORITHM The solution process for a first order linear differential equation is as follows: 1. Put the differential equation in the correct initial form (2.16). x , using (2.19). 2. Find the integrating factor, 3. Multiply everything in the differential equation by becomes the product rule 4. 5. x and verify that the left side y and write it as such. Integrate both sides; make sure you properly deal with the constant of integration. Solve for the solution y . Let’s work a couple of examples. Example 2.8 1. Find the solution of dy 3y 5x such that y1 2 dx x Solution: First, we rearrange the equation to put tin the standard form (2.16) of a linear equation: dy 3 y 5x dx x Multiply both sides of this by the integrating factor: x e px dx e To give the equivalent equation: x 3 3x dx e3 log x x3 dy 3x 2 y 5 x 4 dx 59 d 3 x y , so by integrating both sides, we get: dx Notice that the left – hand side is x 3 y 5x 4 dx x 5 c y x5 cx 3 Now, to satisfy the initial condition, y1 2 , the value of c must be 1. Therefore, the solution to the IVP is: y x 5 x 3 . 2. Solve the differential equation xy 2 y 4 x 2 Solution: y 2x y 4 x …………………………..(*) Integrating factor, x e x 2 dx 2 Multiply (*) through by x : 2 e 2 ln x eln x x 2 . x 2 y 2 xy 4 x3 x 2 y 4x3 x 2 y 4 x3dx x 4 c 3-2.3 Equations Reducible to Linear Form Other first order equations in first degree which are not linear may be reduced to the linear form by means of appropriate transformations. No general rule can be stated; in each instance, the proper transformation is suggested by the form of the equation. Examples of such equations are Bernoulli’s equation, Ricatti’s equation and the one considered below. Consider xs y t y y y ……………………………. (2.21) in which the coefficient of y is a linear function of x and No other x appears anywhere again. Thus (2.21) can be expressed as a linear differential equation with x as the dependent variable and y the independent variable. To be able to do this, multiply (30) through by to obtain: y dx dy dx dx xs y t y or y s y x t y dy dy Thus this is a linear differential equation of x as a function of the independent variable y . 60 Example 2.9 Solve the differential equation 3x 4 y 3 y y 0 . Solution: Clearly the coefficient of y is linear in x and further more x appears nowhere again. dx to obtain: dy dx dx 3x 4 y 3 y 0 y 3 x 4 y 3 dy dy dx 3 x 4y2 Whose standard form is : dy y Therefore we multiply through by Integrating factor: y e dy 3 y d xy 3 3 e ln y y 3 y 3 dx 3y 2 x 4 y5 dy 4 y5 xy3 4 / 6y 6 c 3xy 3 2 y 6 c BERNOULLI’S EQUATION A Bernoulli equation is a first order differential of the form: y pxy qxy n …………… (2.22) in which n is any real number. You would notice that the equation differs from a standardized first order linear differential equation in first degree at the right side of the equation. Here there is a multiplication of y n , n 0, n 1 , at the right hand side(RHS) Then divide (2.22) through by y n and substitute z y1 n . yy n pxy1 n qx Let z y1 n z 1 ny n y The equation becomes: z y n y 1 n z 1 n px z 1 n qx The solution is obtained using first order linear equation in first degree solution techniques. Example 2.10 Solve the differential equation x 2 y x3 dy y 4 cos x . dx 61 Solution: Write it in the form y pxy qxy n y i.e. Divide through by Let z y 3 We get: y y cos x 3 y4 x x 3 : y 4 (I) y 3 cos x y 3 x x z 3 y 4 y , and substitute into (I) 3 3 cos x , which is linear first order in first degree. z z x x3 and x e p x dx Therefore, But …. zy z x3 3 e 3 cos x 3 x dx x3 3x dx e3 ln x x3 3 cos xdx zx3 sin x c x3 3 3 sin x c y x3 y 3 sin x c 3 Hence, Exercise 2.6 Solve the following differential equation. 1. 2. 3. 4. 5. xy 3 y 6 x3 Ans y x3 x3c 1 1 y ce x sin x cos x y y sin x 2 2 dQ 2 4t 2 40t 1304 Q 4; Q 2 100 Q t t 2 dt 10 2t 2t 10 1 y x y y y 2 ce 1 2 1 31 8 y y x9 y 5 ; y 1 2 x 2 x10 4 x y 16 62 RICCATI’S EQUATION The equation dy Px Qx y Rx y 2 ………… (2.23) is non – linear, and needs a dx particular solution, before one can have the chance to solve it. Consider the new function z defined by: z 1 , where y1 is the particular solution. y y1 dz Qx 2 y1Rx z Rx …… (2.24) dx Which is a linear equation satisfied by new function z . Once it is solved, we go back to via 1 the relation: y y1 . z Then easy calculations give: Example 2.11 Solve the equation dy 2 y y 2 where y1 2 is a particular solution. dx Solution: First we verify that y1 2 is a solution, otherwise our calculation is fruitless. 1 1 y 2 z y2 z y 2 z Hence, from the equation satisfied by y , we get: Let z z 3 1 1 1 2 2 2 2 2 2 z z z z z z z 2 Hence z 3z 1 . This is a linear equation. The general solution is given by: 1 1 z e 3x e3x c ce3x 3 3 1 Therefore, we have: y 2 13 ce 3x 63 Note: If one remembers the equation satisfied by z , then the linear equation satisfied by the dz Qx 2 y1Rx z Rx 1 4z 1 3z 1 , since dz Px 2, Qx 1, and Rx 1. functions z is: Exercise 2.7 dy 2 cos2 x sin 2 x y 2 and solve the dx 2 cosx differential equation using the initial condition y0 1 1. Verify that y1 sin x is a solution to 2. Solve the following; i) y 1 x y 2 2 x 1y x ; given solution y 1 ii) y y 2 xy 1; given solution y x iii) y 8xy 2 4 x4 x 1 y 8x3 4 x 2 1 ; given soln. y x 64 Session 4-2 Applications of First-Order Differential Equation (Modeling) We now move into one of the main applications of differential equations. Modeling is the process of writing a differential equation to describe a physical situation. Almost all of the differential equations that you will use in your job (as an engineer) are there because somebody, at some time, modeled a situation to come up with the differential equation that you are using. This session is not intended to teach you how to go about modeling physical situations. A whole course could be devoted to the subject of modeling and still not cover everything! This session is designed to introduce you to the process of modeling and show you what is involved in modeling. In all of these situations we will be forced to make assumptions that do not accurately depict reality in most cases, but without them the problems would be very difficult and beyond the scope of this discussion (and the course in most cases to be honest). We will look at different situations in this session: Law of exponential change – that deals with growth and decay problems – Temperature problems, Falling Bodies, Mixing Problems and Electrical Circuits. But we will part particular attention on electrical problems. 4-2.7 Electrical Circuits Fig. 2.2 The diagram in Figure 2.2 represents an electrical circuit whose total resistance is a constant R ohms and whose self-inductance, show as a coil, is L henries, also a constant. There is a switch whose terminals at a and b can be closed to connect a constant electrical source (may be emf) of E (or V ) volts. Ohm’s law, E IR or V IR , has to be modified for such a circuit. The modified form is: L dI dt RI E …………………..(2.38) 65 where I is the intensity of the current in amperes and t is the time in seconds. By solving this equation, we can predict how the current will flow after the switch is closed. For an RC circuit consisting of a resistance, a capacitance C (in farads), an emf and no inductance (Fig 2.2) the equation governing the amount of electrical charge q (in coulombs) on the capacitor is: dq 1 dt RC The relationship between q and I dq dt q I E R is: ………………..(2.39) …………………………(2.40) Example 1. The switch in the RL circuit in (Fig 2.2) is closed at time t 0 . How will the current flow as a function of time? Solution: Equation (2.38) is a liner first-order differential equation for standard form is: dI R I I as a function of t . Its E ……………………(2.41) dt L L and the corresponding solution, given that I 0 when t 0 , is: E E R/ Lt I e ……………………(2.42) R R Since R and L are positive, R / L is negative and e R / Lt 0 as t . Thus, lim I t E E E E E lim e R/ Lt 0 t R R R R R At any given time, the current is theoretically less than E / R , but as time passes, the current approaches the steady-state value E / R . According to the equation: dI RI E , dt I E / R is the current that will flow in the circuit if either L 0 (no inductance) or dI 0 (steady current, I constant). dt L Equation (2.42) expresses the solution of equation (2.41) as the sum of two terms: a R / Lt steady-state solution V / R and a transient solution V / R e that tends to zero as t . 66 An RL circuit has emf of 5 volts, a resistance of 50 ohms, an inductance of 1 henry, and no initial current. Find the current in the circuit at any time t 2. Solution: Here E 5, R 50 , and L 1, hence (2.41) becomes: The corresponding general solution is: I ce 50t dI dt 50 I 5 101 Using the initial condition I 0 when t 0 , we get I 101 e50t 101 ………………….(I) as a particular solution. 1 e The quantity 10 50t in (I) is called the transient current, since this quantity goes to zero (“dies out”) as t . The quantity As t , the current I 1 10 in (I) is called the steady – state current. approaches the value of the steady – state current. The RL circuit has an emf given (in volts) by 3sint , a resistance of 10 ohms, an inductance of 0.5 henry, and an initial current of 6 amperes. Find the current in the circuit at any time t 3. Solution: Here, E 3sin 2t , R 10 , and L 0.5 ; hence (2.41) becomes: General solution is: I ce At t 0 , I 6 ; 20t dt 20I 6sin 2t 30 3 101 sin 2t 101 cos2t 30 3 6 ce200 101 sin2 0 101 cos2 0 c The current at any time t is: I The steady state current: 4. dI 609 101 609 101 30 3 e20t 101 sin 2t 101 cos2t 30 3 I 101 sin 2t 101 cos2t The RC circuit has an emf given (in volts) by 400sint , a resistance of 100 2 ohms, an capacitance of 10 farad. Initially there is no charge on the capacitor. Find the current in the circuit at any time t Solution: We first fine the charge q and then use (2.40) to obtain the current. Here, E 400cos 2t , R 100 , and C 102 ; hence (2.39) becomes: 67 dq q 4cos 2t dt This equation is linear, and its solution is (two integrations by parts are required) q cet 85 sin 2t 54 cos2t . When there is no charge, that means, q 0 , so at t 0, q 0 ; hence 0 ce 0 85 sin 2 0 54 cos 0 c 54 Thus, q 54 e t 85 sin 2t 54 cos 2t and using (2.40), we obtain: I dq dt t 54 e 165 sin 2t 85 cos 2t Summary - General first – order differential equation: F x, y, y 0 - Standard form: - Differential form: - First – order differential equation is said to be separable, if the differential form have y f x, y M x, y dx N x, y dy 0 M x, y m x p y , which the product of a function x and a function y and also N x, y n x q y , also a function of x and a function, then we can write the differential form as: m x p y dx n x q y dy 0 We then divide throughout by p y n x , we get: m x p y dx dy 0 n x q y now it is separated in each variable x and y . Integrating both sides: m x p y dx dy c n x q y where - c is arbitrary constant. Separation of Variables. To solve the initial – value problem y f x, y ; y x0 y0 68 (a) Separate the variables, so that all of the x -dependence is on one side of the equation, and all of the y -dependence is on the other side of the equation. (b) (c) Integrate both sides of the equation. Substitute the initial condition to solve for the constant of integration - Reduction to Separable o A formal definition of homogeneity is: A function of two variables, f x, y , is said to be homogeneous of degree n if there is constant n such that: f x, y n f x, y o A differential equation of the form (differential form): M x, y dx N x, y dy 0 is homogeneous if the functions M x, y and N x, y are both homogeneous functions of the same degree. Or If the equation in standard form: y f x, y depends only on the ratio y x or , then it is said to be homogenous. x y To solve homogeneous equations, turn them into separable ones using the substitution: y x , where is a function of x ; also: dy d x or dy xd dx dx dx o A Transformation (Equations in which N x, y are linear but not homogeneous) M x, y and Another type of equations that can be reduced to a more basic type by means of a suitable transformation is an equation of the form: a1x b1 y c1 dx a2 x b2 y c2 dy 0 where a1 , b1, c1 , a2 , b2 , c2 are constants where M x, y a1 x b1 y c1 and N x, y a2 x b2 y c2 written as M x, y N x, y y 0 , which are not homogeneous. However the differential equation can be made homogeneous and solved accordingly provided; 69 Case I a2 b2 , then the transformation: a1 b1 x X h and y Y k , where h, k is the solution of the system a1 x b1 y c1 0 and a2 x b2 y c2 0 , that is x h and y k , which reduces to a homogeneous equation: a1X b1YdX a2X b2Ydy 0 in the variables x and Y . If Case II a2 b2 k , then the transformation z a1x b1 y reduces the a1 b1 equation to a separable equation in the variables x and z : If dz dy a1 b1 dx dx - Exact Equations Given a function f x, y , its total differentials, df , is defined as: df f f dx dy x y This shows that the family of curves (or general solution) f x, y c satisfies the differential equation df 0 . f f dx dy 0 x y So if there exists a function f x, y such that: M x, y f f and N x, y y x then M x, y dx N x, y dy is called an exact differential, and the equation: M x, y dx N x, y dy 0 is said to be an exact equation, whose solution is the family f x, y c . - Integrating Factors – Non-exact If M x, y dx N x, y dy 0 is not exact, sometimes we can turn it into an exact differential equation by multiplying the whole equation by an appropriate factor, called an integrating factor, let say x, y , which is a function of x, y . 70 For which we obtain: M x, y dx N x, y dy 0 and expect it to be exact. M N For it to be exact then: y Applying product rule we get: x M N M N y y x x N M N y x x y M y N x N x M y M Simply put: N x M y M y N x o Linear First – Order Equations A first – order linear in the standard form is written as y px y qx ………………………………(2.16) We recognise (2.16) as an inexact differential equation which requires an integrating factor x , that is: d x 0 dy x to obtain the following: xyx pxyx qx 0 ……………………..(2.17) M x, y x px y qx and N x, y x Multiplying through (2.16) by M N . y x py q x d x Thus, y x dx d p ………….(2.18) dx d pdx ln pdx Separating variables: If (2.17) is exact, then e pdx ……………………..(2.19) Without loose of operating the constant of integration of (2.19) is unity. Consider (2.16) and multiply through by: x e then we have the exact equation: y py q y y q by (2.18) that is p 71 p x dx , d p p dx Integrating both sides: y qdx c y 1 qdx 1c p x dx p x dx p x dx ye qxdx ce ……(2.20) e The solution is in two parts in the R.H.S. The term ce p x dx 1c , called the complementary function is the solution of the reduced equation: y p x y 0 obtained by setting qx 0 in (2.16). The differential equation which has qx 0 is called Homogeneous differential equation. The other term on the R.H.S of the solution is dependent on qx , that 1 qdx is called the particular integral and it satisfies (with the complementary function) the differential equation: y p x y q x , called the Non – Homogeneous differential equation. Notice that the particular integral contains no constant. o Bernoulli’s Equation A Bernoulli’s equation is a first order differential of the form: y pxy qxy n ……………(2.22) in which n is any real number. Here there is a multiply of y n , n 0, n 1 , at the right hand side(RHS) n Then divide (2.22) through by y and substitute yy n pxy1 n qx Let z y1 n z 1 ny n y z y n y 1 n The equation becomes: z 1 n px z 1 n qx 72 z y1 n . The solution is obtained using first order linear equation in first degree solution techniques. o Riccati’s Equation dy Px Qx y Rx y 2 dx Consider the new function z defined by: 1 z y y1 The equation where y1 is the particular solution. dz Qx 2 y1Rx z Rx dx Which is a linear equation satisfied by new function z . Once it is solved, we 1 go back to via the relation: y y1 . z Then easy calculations give: 73 UNIT 3 LINEAR DIFFERENTIAL EQUATION OF HIGHER ORDER Introduction So far, all equations we have studied have been first order, which means that they contained only the first derivative, say y . Now we will turn to higher – order equation. We will start with one of the most common (and, fortunately, easily solved) types of higher – order equation – the linear equations with constant coefficients then we proceed with other forms. Learning Objectives After going through this unit, you would be able to: • solve most linear differential equation of higher order with constant coefficient. • explain the method of attack to a particular linear differential equation and which method is most efficient • recognise that solutions of higher order differential equation are linearly independent and can use the Wronskian to determine it. Session 1-3 Linear Differential Equation of Higher Order A general n th – order linear differential equation has the form: dny d n1 y d2y dy a x ... a x a1 x a0 x y R x ….. (3.1) n1 2 n n1 2 dx dx dx dx where an x 0 and the coefficients ai x i 0,1,...n 1, n ,and f x are given functions of x or an x are constants. This equation (3.1) is homogeneous if f x 0 and non-homogeneous if f x 0 . The solution of the homogeneous differential equation is called the complimentary function (conventionally yc ) and the solution of the non-homogeneous part is known as the particular integral (conventionally y p ). Then the general solution = complimentary function + particular integral (that is y yc y p ). 74 1-3.1 Homogeneous Equation with Constant Coefficients A general homogeneous differential equation with constant coefficients is: an y n an1 y n1 ... a1 y a0 y 0 …..(3.2) We start problem solving with second order differential equation of the form: 2 a2 d y dx 2 a1 dy a0 y 0 ……………………... dx (3.3) We consider a trial solution y e , exponential function, suitably chosen r to be determined rx for the constant coefficient linear differential equation. Then (3.3) becomes: a r 2 clearly 2 a1r a0 e 0 …………………………….(3.4) erx 0 , hence; rx a2 r a1r a0 0 ……………………….(3.5) 2 or for convenience, ar br c 0 ……………………………(3.5) 2 Hence if r is root of this quadratic equation (3.5), often called the auxiliary or characteristic rx equations, e is a solution of (3.3). b b 4ac 2 The roots of (3.5) are: r1 2a There are three cases to consider: b b 4ac 2 and r2 2a Case 1: If the roots are real and distinct b 4ac 0 , that is, r r1 , 2 r r2 and r1 r2 then the general solution of the differential equation is: y c1e r1 x c2er2 x Case 2: If the roots are real and identical b 4ac 0 , that is, r r1 r2 , if we denote the 2 double root by r , then the general solution of the differential equation is: y c1e c2 xe rx rx Case 3: If the roots are not real. In this case, the roots are complex conjugates b 2 4ac 0 , that is r i , then the general solution of the differential x equation is y e c cos x c sin x where 1 2 Example 3.1 1. Solve the differential equation y 2 y y 0 75 b 2a b 4ac 2 ; 2a Solution: Let y e , y re , y r e : rx rx 2 rx r e 2re e 0 r 2r 1 0 : r 1 x x y c1e c2 xe 2 rx 2. rx rx 2 Solve the differential equation y 2 y 5 y 0 . Solution: Aux: r 2r 5 0 r 1 2i 2 ye 3. x c cos 2x c sin 2x 1 2 d 2 y dy 6y 0 2 dx dx Solve Solution: Aux: r r 6 0 r1 2, r2 3 2 y c1e c2e 2x 3 x 1-3.2 Homogeneous with constant coefficient of order 3 and higher Theorem I: Distinct Real Roots In case 1; for an n th order linear differential equation with constant coefficients, if r1 , r2 ,...rn are distinct roots of the characteristic equation resulting from the differential equation, then the general solution is of the form: y c1e 1 c2e 2 .... cne n r r r Theorem II: Repeated Roots In case 2; if the characteristic equation has a repeated root r of multiplicity k , then the part of a general solution of the differential equation of the form: y c1 c2 x c3 x ... ck x 2 k 1 e rx Example 3.2 Find a general solution of y 1. 3 y 6 y 0 Solution: Aux: r r 6r 0 r r 3 r 2 0 ; r 0, r 3, r 2 3 2 Hence, y c1 c2e c3e 3x 2. 2 x Find a general solution of y 4 3y 3 76 3 y y 0 Solution: Aux: r 3r 3r r 0 4 3 2 r1 0, and it has also the triple k 3 root r2 1 y c1 c2 c3 x c4 x e 2 x Exercise 3.1 Solve the following linear differential equation. 1. y 4 y 0 Ans. y c1 cos 2 x c2 sin 2 x 2. 4 y 4 y 3 y 0 y c1e 2 c2e 3. 2 y 9 y 0 y c1 c2e 2 4. 9 y 12 y 4 y 0 y c1 c2e 2 x / 3 c3 xe 2 x / 3 5. y y y y 0 6. y 4 y 4 y 0; y 0 1, y 0 1 y c1e c2e 7. y 2.2 y 1.17 y 0; y 0 2, y 0 2.6 y 0.3e 8. y 6 y 25 y 0; y 0 3, y 0 1 ye 9. y 4 4 y 0 10. y 10 y 25 y 0; y 0 3, y 0 4, y 0 5 x 23 x 9x x y 1 3x e 3x x /4 x c3 xe x 2 x 0.5e x /2 3cos 4 x 2sin 4 x y c1 c2 x c3 cos 2 x c4 sin 2 x 1-3.3 Theory of Solutions of Linear Differential Equations Consider a Homogeneous second – order linear differential equation y px y qx y 0 ……………………………(3.6) Principle of Superposition Let y1 and y2 be two solutions of the homogeneous linear equation (3.6) on the interval I . If c1 and c2 are constants, then the linear combination y c1 y1 c2 y2 ………………………(3.7) is also a solution of (3.6) on I . Prove as exercise. 77 Definition: Linear Independence of Two Functions and the Wronskian Two functions defined on an open interval are called linearly independent provided that neither is a constant multiple of the other. Two functions are said to be linearly dependent if they are not linearly independent; that is, one is a constant multiple of the other. In general, a set of functions (solutions to differential equation) y1 x , y2 x ,..., yn x is linearly dependent on a x b if there exist constants, c1 , c2 ,..., cn , not all zero, such that c1 y1 c2 y2 ... cn yn 0 ………. (3.8) on a x b . Example 3.2 y1 cosx and y2 sin x are two solutions of the equation y y 0 . By principle of superposition y 3 y1 2 y2 3cosx 2 sin x is also a solution. We can always determine whether two given functions f and g are linearly dependent on an interval I by noting at a glance whether either of the two quotients f / g and g / f as a constant on I . The following pairs are linearly independent on the entire real line: sin x,cos x , e x , e 2 x , e x , xe x , x 1, x 2 , and x, x . Given two functions f and g , the Wronskian of f and g is the determinant : W f , g f g fg f g f g For example, W cosx, sin x ex while W e , xe x e x x cosx sin x cos2 x sin 2 x 1 0 sin x cosx xe x e 2 x 0x x x e xe Given two functions f x and g x that are differentiable on some interval I : 1. 2. If W f , g x 0 for some x0 in I , then f x and g x are linearly independent on the interval F . If f x and g x are linearly dependent on I then W f , g x 0 for all x is the interval. 78 Note: It is possible for two linearly independent functions to have a zero Wronskian. Suppose that y1 and y2 are two solutions of homogeneous second order linear equation y px y qx y 0 on open interval I on which px and qx are continuous. If W y1, y2 0 , then the two solutions are called a fundamental set of solutions and the general solution to (3.6) is y c1 y1 c2 y2 . Theorem 3.1 The nth - order linear homogeneous differential equation (3.2) always has n linearly independent solutions. If y1 x , y2 x ,..., yn x represent these solution, then the general solution (3.2) is y x c1 y1 x c2 y2 x cn yn x …… (3.9) where c1, c2 ,..., cn denote arbitrary constants. Exercise 3.2 Determine if the following sets of functions are linearly dependent or linearly dependent a) f x 9 cos 2 x , g x 2 cos2 x 2 sin x b) f x 2x 2 , g x x 4 1-3.4 Reduction of Order Consider the non – constant coefficient, second order differential equations of the form y px y qx y 0 ………………(3.6) In general, finding solutions to these kinds of differential equations can be much more difficult than finding solutions to constant coefficient differential equation. We discuss here the method of reduction of order, which enables us to use one known solution, say, y1 to find a second linearly independent solution y2 . Solution of Reduction Order: y px y qx y 0 ……………… (3.6) Consider Let y2 x vxy1x………………………. (3.10) 79 We begin by substituting the expression in (3.10) into (3.6), using the derivatives: y2 vy1 vy1 y2 vy1 2vy1 vy1 We get: vy1 2vy1 vy1 pvy1 vy1 qvy1 0 and rearrangement gives vy1 py1 qy1 vy1 2vy1 pvy1 0 but y1 py1 qy1 0 is y1 is a solution of (3.6). This leaves the equation y1v 2 y1 py1 v 0 …………. (3.11) The key to the success of this method is that (3.11) is linear in v . Thus the substitution of (3.10) has reduced the second order linear equation in (3.6) to the first order (in v ) in linear equation in (3.11) If we write u v and assume that y1 never vanishes on I , then (3.10) yields: 2 y u 1 p( x) u 0 ……………………………… (3.12) y1 An integrating factor for (3.9) is: y exp 2 1 px dx exp2 ln y1 y1 y12 e p Thus x dx 2 We now integrate the equation in (3.6) to obtain: uy1 e So v u p ( x) dx p( x)dx c p x dx e 2 c y1 p x dx y2 e vc dx k Another integration now gives: y1 y12 With the particular choices c 1 and k 0 , we get: p x dx e y2 y1 dx ………………(3.13) y12 Example 3.3 1. 1 2 Find the general solution to 2 x y xy 3 y 0 given that y1 ( x) x is a solution. Solution: 80 2x 2 y xy 3 y 0 y y2 y1 p x dx e y12 1 1 3 y 2 y 0 where p( x) . 2x 2x 2x dx y2 x 1 1 dx e 2x x 1 x 2 dx x 2 dx 2 x 1 1 3 3 5 y2 x 1 x 2 dx x 1 52 x 2 c let c 0 y2 2 x 2 5 3 y c1 x 1 c2 x 2 Thus a general solution is: Given the solution y1 x , find a general solution of 4. x 2 2 1 y 2 xy 2 y 0 (x 1) Solution: 2 First divide each term by x 1 to obtain: y With px 2x / x 2 1 . p x dx exp Hence e 2x x 1 2 y 2 x 1 2 y0 dx exp ln x 2 1 1 x 2 since x 2 1 . x 1 2x 2 Therefore, the formula in (37) yields: y2 x Thus a general solution is: 1 x2 x y c1 x c2 1 x 2 2 dx x x 2 1dx 1 x 2 Exercise 3.3 Given one solution y1 for each question, find a second linearly independent solution y2 of the following differential equation: 1. y 4 y 4 y 0; 2. x 2 y 3xy y 0; y1 1x 3. x 2 y 4xy x 2 6 y 0 ; y1 x 2 sin x 4. x 2 y xy 9 y 0; y1 x 3 5. x 1y x 2y y 0; y1 e 2 x x 6. 4 y 4 y y 0 ; y1 e 2 7. xy 3 y 0 ; y1 1 y1 x 3 81 1-3.5 Linear Non – Homogeneous Equation with Constant Coefficients A general linear non – homogeneous constant coefficient equation is: dny d n 1 y dy a0 y f x ………………… (3.14) dx dxn dxn 1 f x is a given function of x , and a0 , a1,...,an 1, an also given constants. an an 1 ... a1 The general solution of (3.11) is given as: General solution = Complimentary Function yc x + Particular Integral y p x There are several different methods for finding solutions of non-homogeneous linear equations. Here we study three methods to find our Particular Integral, since the complimentary is from the homogeneous part, which we have done already; (I) Method of Undetermined Coefficients (II) Method of Variation of Parameter (III) Method of the Linear Differential Operator 1-3.6 Method of Undetermined Coefficients The main focus of this section is the structure of solutions of non-homogeneous linear differential equations rather than solution methods. Knowledge of the structure of solutions makes it easier to find solutions. The method of undetermined coefficients is a method for finding particular solutions for some differential equations of the form (3.14). The method consists of three steps: a. We assume a trial solution function y p x , a family of functions that is guaranteed to include a correct particular solution; Find the image of the trial solution whose form depends on f x of (3.14). b. y p x is then substituted into (3.14) and the constants take values which satisfy c. (3.14) completely (a) bx For f x ae a, b are constants, if the auxiliary equation of (3.14) has r b as a root of multiplicity k , then the trial function is of the form y p x Ax k ebx , where A constant to be determined. If r b is Not a root, take k 0 . Example 3.4 Find a particular integral of y 3 y 2 y e 2 x . Solution: The auxiliary equation is: r 3r 2 0 , r 2 r 1 0 r 2,1 2 82 Hence f x e 2 x that is b 2, a root. The trial function is: y p x Axe 2 x Substituting this into the differential equation: y p x Axe 2 x , yp x Ae 2 x 2 xAe2 x yp x 2 Ae 2 x 2 Ae 2 x 4 Axe2 x 4 Ae2 x 4 Axe2 x 3Ae2 x 6 Axe2 x 2 Axe2 x e 2 x Ae2 x e 2 x A 1 y p x xe 2 x General solution: y yc y p c1e (b) 2x c2e x xe2 x For f x a sin bx or cos bx or both, a, b constants, if r 2 b2 is a factor of the auxiliary equation of multiplicity k , then: y p x x k A sin bx B cos bx , where A, B are constant to be determined. If r b is Not a factor, then k 0 . 2 2 Example 3.4 Solve the differential equation y 4 y 3sin 2 x . Solution: r 2 0, b 2, k 1 y p Ax sin 2 x Bx cos 2 x --------------- (I) 2 2 yp Asin 2 x 2 x cos 2 x Bcos 2 x 2 x sin 2 x ------ (II) yp A2 cos2 x 2cos2 x 2 x sin 2 x B 2 sin 2 x 2sin 2 x 2 x cos2 x Substitute back into the equation gives: --------------- (III) A2 cos 2 x 2 cos 2 x 4 x sin 2 x B 2 sin 2 x 2 sin 2 x 4 x cos 2 x 4Ax sin 2 x Bx cos 2 x 3sin 2 x 4 A cos 2 x 4B sin 2 x 3sin 2 x Comparing the coefficients: A 0, B 3 4 3 y p x x cos2 x 4 83 General solution: yx c1 cos 2 x c2 sin 2 x (c) 3 x cos 2 x . 4 For f x axs , a, s (integer) constant, if the auxiliary equation has r 0 as a root of multiplicity k , then trial function: y p x x k Ax s Bx s 1 Px Q where A, B, , P, Q are constants to be determined. If r 0 is Not a root, then k 0 Example 3.5 Find a particular solution of y 4 y 3x3 . Solution: y p Ax3 Bx 2 Cx D yp 3 Ax 2 2Bx C yp 6 Ax 2B Substituting in the differential equation, we get: y 4 y p 6 Ax 2B 4 Ax3 Bx 2 Cx D 4 Ax3 4Bx2 6 A 4C x 2B D 3x 3 . We equate coefficients of like powers of x to get the equations: 4 A 3, 4B 0 6 A 4C 0, 2B D 0 with solution A 34 , B 0, C 98 , and D 0 . Hence a particular solution of the 3 differential equation is: y p 3 x 9 x . 4 8 If f x is the sum of any two or all these special functions, the particular integral (d) is the appropriate sum of individual particular integral. Hence f x takes the trial solution: y p x k A0 A1 x ... An x n e mx cos bx B0 B1 x ... Bn x n e mx sin bx Example 3.6 Determine the appropriate form of for a particular solution of y 6 y 13 y e 3x cos2x Solution: The characteristic equation r 6r 13 0 has roots: r 3 2i , 2 3 x So the complementary function is: yc e A cos 2 x B sin 2 x 84 This is the same form as a first attempt e 3x A cos2 x B sin 2 x at a particular solution, so we must multiply by x to eliminate duplication. Hence we would take y p e 3x Ax cos 2 x Bx sin 2 x Exercise 3.4 Solve the following linear differential equation. 1. y y 2 y 4x2 y c1e x c2e2 x 2x2 2x 3 2. y y 2 y e3x y c1e x c2e2 x 14 e3 x 3. y y 2 y sin 2 x 4. y 4 y 2e2 x y c1e2 x c2e2 x 12 xe2 x 6. y 6 y 10 y 20 e 2 x y 4 y 3x cos 2 x 7. y 6 y 13 y xe3x sin 2x 8. y 3 y 2 y x e x e 2 x 9. y 2 y 3 y 4e3x 5 5. 10. d3y 3 d2y 2 y c1e x c2e2 x 203 sin 2 x 201 cos2 x 3e x 4x 2 11. dx dx y 6 y 8 y cosh x given y0 0 12. y y 3 2e x 85 y0 1 1-3.7 Method of Variation of Parameters In method (1-3.6), we presented the method of undetermined coefficients which, when it is applicable, is usually the simplest method of finding a particular solution of a nonhomogeneous linear differential equation with constant coefficients. This method (1-3.6) cannot succeed with an equation such as y y tan x , because the function f x tan x has infinitely many linearly independent derivatives. Consider a non-homogeneous linear second order differential equation of the form: y px y qx y f x ……………….. (3.15) Suppose y1 and y2 are the two solutions of the homogeneous equation y px y qx y 0 …………………… (3.16) Then the complimentary function is: yx c1 y1x c2 y2 x ……………………. (3.17) where c1 and c2 are arbitrary constant. To find the particular integral, this method requires the trial function y p x u1 x y1 x u 2 x y2 x …………………….(3.18) where we replace the constants, or parameters c1,c2 in the complementary function in (3.17) by variables: functions u1 and u2 of x . (3.18) is substituted into (3.15) to determine u1x and u2 ( x) . To determine u1 x and u2 x the following conditions should be fulfilled: (I) (II) (3.18) must satisfy (3.15) A condition imposed arbitrarily to facilitate the calculation of the function u1x and u2 x . y p x u1 y1 u 2 y2 ……………………………………..(3.19) y p x u1 y1 u1 y1 u 2 y2 u 2 y2 Let u1 y1 u2 y2 0 ……………………….(3.19) : condition imposed y p x u1 y1 u 2 y2 …………………….. (3.20) y p x u1 y1 u1 y1 u 2 y2 u 2 y2 …. (3.21) Substituting (3.21), (3.20) and (3.18) into (3.15), we have: u1 y1 u1 y1 u2 y2 u2 y2 pu1 y1 pu2 y2 qu1 y1 qu2 y2 f x u1 y1 py1 qy1 u2 y2 py2 qy2 u1 y1 u2 y2 f ( x) 86 But both y1 and y2 satisfy the homogeneous equation (3.16) associated with the nonhomogeneous equation (3.15). So y1 py1 qy1 0 and y2 py2 qy2 0 Equation (3.15) finally becomes: u1 y1 u2 y2 f x ………………….. (3.22) Using (3.19) and (3.22) u1 and u2 can be determined and thus we have a system: u1 y1 u2 y2 0 …………………. (3.23) u1 y1 u2 y2 f x The Cramer’s Rule for the solution of a linear system of equation yields: 0 f x u1 y1 y1 u1 y2 f x W y1, y2 Wronskian. u1 y2 y2 , y2 y2 y1 0 y f x u 2 1 y1 y2 y1 y2 , u2 y2 f x dx , W y1 , y2 u2 Then (3.18) becomes: y p y1 x y1 f x ; since the denominator is the W y1 , y2 y1 f x dx W y1 , y2 y2 x f x y x f x dx y2 x 1 dx ……..(3.24) W y1 , y2 W y1 , y2 Example 3.7 Find the general solution of the differential equation y y sec x 0 x 2 Solution: The two solutions of the homogeneous equation are: y1 cosx and y2 sin x . Then y p u1 x cos x u 2 x sin x y p u1 cos x u1 sin x u2 sin x u 2 cos x Imposition: u1 cosx u2 sin x 0 …………….(I) yp u1 x sin x u2 x cos x y p u1 sin x u1 cos x u 2 cos x u 2 sin x 87 Then differential equation becomes: u1 sin x u1 cosx u2 cosx u2 sin x u1 cosx u2 sin x sec x u1 sin x u2 cosx sec x ………………………….(II) Using (I) and II, we have: u1 cos x u2 sin x 0 u1 sin x u2 cos x sec x The Cramer’s Rule for the solution of a linear system of equation yields: 0 sec x u1 cos x sin x sin x cos x , sin x cos x cos x 0 sin x sec x u2 cos x sin x sin x cos x u1 tan x , u2 1 u1 ln cosx , u2 x Therefore, y p ln cos x cos x x sin x . Hence the General solution: yx c1 cosx c2 sin x ln cosxcosx x sin x Exercise 3.5 Determine a particular solution of the following using the method of variation of parameter. 1. y 5 y 6 y e 2 x 9. y 3 y 2 y e x e 2 x 2. y y 2 y 2e x 10. y y cos x 3. y 2 y y 3e x 11. y y cot x 4. y y tan x 12. y 5 y 6 y e 2 x e3x 5. y 9 y 9 sec2 3x 0 x 6 13. y y sin x 6. y 4 y 3 cos ec2x 0 x 2 7. y 3 y 2 y e x 8. y 2 y y e x 0 x 2 88 Session 2-3 Linear Differential Operator Method (D-Operator) A differential operator is a rule D that uses derivatives to assign a function D y to any sufficiently differentiable function y . The function D y is called the image of y under D . This method is very efficient for the solution of a general linear equation of n th order and with constant coefficients. It is most efficient because it can be manipulated algebraically. 2 Let D denote differentiation with respect to x , D differentiation twice with respect to x , k Then D( x 2 ) 2x d y d , that is D k dx dx D2 ( x3 4x 2 1) 6x 8 k and so on; that is, for positive integral k , D y D2 (e3x ) 9e3x The expression f D a0 D a1D n1 d2 D 2 etc dx 2 an1D an is called a differential operator of order n . It may be defined as that operator which, when applied to any function y , yields n the result f D y a0 n d y a1 d n1 y an1 dy an y n n1 dx dx dx where the function y is assumed to possess as many derivatives as may be encountered in whatever operation take place. Let f D be a polynomial in D . 2-3.1 Fundamental Properties of the D–operator The following properties are satisfied by the D-operator. D n (u v) D n u D n v (Distributive property) Dn (ku) kDnu , k is a constant D m (D n u) D mn (u) 2-3.2 Factorability Any constant coefficient n th order linear differential equation (LDE) can be written as: f ( D)y R( x) , where f (D) is a polynomial function of the operator D . Example: y 3y 2 y D2 y 3Dy 2 y e 2 x ( D 2 3D 2) y e 2 x f (D) D2 3D 2 This polynomial can be factorized if required. In this case f (D) D2 3D 2 (D 2)(D 1) 89 We recognize that (D 2)(D 1) (D 1)(D 2) which is valid ONLY for constant coefficient linear equations. 2-3.3 Linear Equations of First Order Recall the familiar first-order linear equation with constant coefficient: dy ay R(x) , a is dx a constant, which is equivalent to: (D a) y R( x) y 1 R( x ) Da The integrating factor is I e e ax . Thus multiplying both sides by the integrating factor we obtain dy e ax aeax y e ax R(x) dx adx d ye ax e ax R( x) dx ye ax e ax R( x)dx y eax e ax R( x)dx (1.1) 1 1 R( x) R( x) e ax e ax R( x)dx Da Da 1 e2x Therefore if we have ( D 2) y e 2 x y D2 From ( ) y 2x 2x 2 x 2x dx e2 x x c xe2 x ce2 x then y e e e dx y e 2-3.4 Linear Equation of Higher Order Here we discuss second order linear equations, aD2 bD c y f ( x) Example 1: Solve D 2 3D 2 y e 2 x Solution D 2 3D 2 y D 2D 1y e 2 x Let D 1y z , then D 2z e 2 x 1 2x e e 2 x e 2 x e 2 x dx e 2 x x c1 D2 2x Thus D 1y e x c1 1 y e 2 x x c1 e 2 x e x e 2 x ( x c1 ) dx D 1 y xe2 x Ae 2 x Be x , A and B are arbitrary constants. z 90 Alternatively; D 1D 2y e2x y 1 e2x D 1D 2 Using partial fraction: 1 2x 1 2x 1 2x 1 y e e e D2 D 1 D 2 D 1 y e 2 x e 2 x e 2 x dx e x e x e 2 x dx xe 2 x c1e 2 x e 2 x c2 e x y xe 2 x (c1 1)e 2 x c2 e x xe 2 x Ae 2 x B2 e x The efficiency of the operator method becomes very clear where the auxiliary equation has repeated roots. Example 2: Solve D2 2D 1 y 2xex Solution: Let z D 1 D 12 y 2xex 1 2 xe x e x e x 2 xe x dx e x x 2 c1 D 1 D 1y e x ( x 2 c1 ) 1 y e x ( x 2 c1 ) e x e x e x x 2 c1 dx D 1 Then D 1z 2xex z ye x (x 2 c1 )dx e x 13 x3 x 3e x c1x c2 c1xe x c2e x 3 Exercise 3.6 1. 2. Demonstrate Commutation of the factor in the operator D 2D 3 applied onto: i) ii) e2x x3 ln x x 3 Use equation 1.1 to solve the following; i) ii) D 2y e x D 1y e x iii) v) vii) D 2y x 2 e2x 2D 3y sin x 3D 2y cos 23 x iv) vi) viii) 91 D 2y x 2 e2x 2D 3y cos x 3D 2y sin 23 x Solve D D i) 2 9 y e3x 5D 6 y e y 7 y 12y x iii) v) 2 D 32 y e 3x D 32 y e3x D 22 y xe3x vii) ix) xi) ii) 3x iv) vi) viii) x) xii) D D 2 9 y e 3x 5D 6 y e 2 x y 9 y 2 y x 2 D 32 y e3x D 32 y e 3x D 32 y xe3x 2-3.5 f D Polynomials with complex roots Consider a differential equation f D y R x for which the auxiliary equation f m 0 has real coefficients. From elementary algebra we know that if the auxiliary equation has any imaginary roots, those roots must occur in conjugate pairs. Thus, if m1 i is a root of the homogeneous equation f m 0 , with a and b real and b 0 , then m2 i is also a root of f m 0 . For aD 2 bD c y R x if f m 0 have complex roots then we have: D i D i y R x Let z D i y then D i z R x . Example 3: Solve D 2 4 y 3 Solution D 2i D 2i y 3 Let D 2i y z zD 2i 3 1 3 3 e 2ix 3e 2ix dx e 2ix e 2ix c D 2i 2i 3 3 3 3 z ce 2ix c1e 2ix , where c1 c 2i 2i 2i 2i 1 3 2ix 2ix 2ix 3 2ix y c e e e c1e dx 1 D 2i 2i 2i z 92 c 3 3 c y e 2ix e 2ix 1 e 4ix c2 1 e 2ix c2e 2ix 4i 4 4 4i c 3 y c3e 2ix c2 e 2ix where c3 1 4 4i y 3 A cos 2 x B sin 2 x 4 Exercise 3.7 1. 3. 5. 7. 9. D( D 1) y 5 D D D D 2 D 1 y 2 4. 2 4 y e2x 6. 2 2 9y sin 2x 2D 1y e D 4y 0 D 4y cos x D 1y e D 2D 1y e sin x 2 2. 8. x 2 2 2x 2 x cos x A general n th – order linear differential equation has the form: dny d n1 y d2y dy an x n an1 x n1 ... a2 x 2 a1 x a0 x y f x ….. (3.1) dx dx dx dx where an x 0 and the coefficients ai x i 0,1,...n 1, n ,and f x are given functions of x or are constants. This equation (3.1) is homogeneous if f x 0 and non – homogeneous if f x 0 . The solution of the homogeneous differential equation is called the complimentary function (conventionally yc ) and the solution of the non – homogeneous part is known as the particular integral (conventionally y p ). Then the general solution = complimentary function + particular integral (that is y yc y p ). - Homogeneous Equation of Second – Order with constant coefficient d2y dy a2 2 a1 a0 y 0 dx dx rx A trial solution y e substituted gives: a2r 2 a1r a0 erx 0 93 clearly erx 0 , hence; a2 r 2 a1r a0 0 or for convenience, ar 2 br c 0 , where a a2 ; b a1; c a0 whose roots: b b2 4ac b b2 4ac r1 and r2 2a 2a There are three cases to consider: Case 1: If the roots are real and distinct b 2 4ac 0 , that is, r r1 , r r2 and r1 r2 then the general solution of the differential equation is: y c1e r1 x c2er2 x Case 2: If the roots are real and identical b 2 4ac 0 , that is, r r1 r2 , if we denote the double root by r , then the general solution of the differential equation is: y c1e rx c2 xe rx Case 3: conjugates b If the roots are not real. In this case, the roots are complex 2 4ac 0 , that is r i , then the general solution of the differential equation is: y ex c1 cos x c2 sin x where - b2 4ac b ; 2a 2a Wronskian Given two functions f and g , the Wronskian of f and g is the determinant : W f , g f g fg f g f g Given two functions f x and g x that are differentiable on some interval I: i. If W f , g x 0 for some x0 in I , then f x and g x are linearly independent on the interval F . ii. If f x and g x are linearly dependent on I then W f , g x 0 for all x is the interval. 94 - Reduction Order In reduction of order, it enables us to use one known solution , say, y1 to find a second linearly independent solution y2 . Solution of Reduction Order: y px y qx y 0 Consider Then y2 y1 - p x dx e y12 dx Second – Order Non – Homogeneous Equation A general linear non – homogeneous constant coefficient equation is: an dny n dx an 1 d n 1 y n 1 dx ... a1 dy a0 y f x dx We have three Methods of attack. Each depend on the external function, f x . The methods are: (IV) (V) (VI) Method of Undetermined Coefficients Method of Variation of Parameter Method of the Linear Differential Operator Method of Undetermined Coefficients (e) For f x aebx a, b are constants, if the auxiliary equation of has r b as a root of multiplicity k , then the trial function is of the form y p x Ax k ebx , where A constant to be determined. If r b is Not is a root, take k 0 . (b) For f x a sin bx or cos bx or both, a, b constants, if r 2 b2 is a factor of the auxiliary equation of multiplicity k , then: y p x x k A sin bx B cos bx where A, B are constant to be determined. If (c) r 2 b2 is Not a factor, then k 0 . For f x axs , a, s (integer) constant, if the auxiliary equation has r 0 as a root of multiplicity k , then trial function: y p x x k Ax s Bx s 1 Px Q where A, B, , P, Q are constants to be determined. If r 0 is Not a root, then k 0 95 (d) If f x is the sum of any two or all these special functions, the particular integral is the appropriate sum of individual particular integral. Hence f x takes the trial solution: y p x k A0 A1 x ... An x n e mx cos bx B0 B1 x ... Bn x n e mx sin bx Method of Variation of Parameter The method of undetermined coefficient does not succeed if the external function, f x is not analytic everywhere. The parameters of the homogeneous solution are then varied (that is change to variables). py qy f x , The homogeneous part, y py qy 0 , gives the solution yc c1 y1 c2 y2 . If the parameters c1 and c2 is varied to u1 ( x) and u2 x Given a differential equation: y respectively, we have: y p u1 y1 u2 y2 We then determined u1 and u2 to get: y p y1 x y2 x f x y x f x dx y2 x 1 dx W y1 , y2 W y1 , y2 Method Linear Differential Operator dy ay R(x) , a is a constant, dx which is equivalent to: (D a) y R( x) 1 y R( x ) Da adx The integrating factor I e e ax Let y eax eax R( x)dx Then from 1 R( x) Da 1 R( x) e ax e ax R( x)dx Da y 96 UNIT FOUR LAPLACE TRANSFORM 4.0 INTRODUCTION In this section we will be looking at how to use Laplace transforms to solve differential equations. Laplace transforms reduce a differential equation to an algebra problem. We use Laplace transforms when the force function becomes complicated, that is, discontinuous – for instance, when the voltage supplied to an electrical circuit is periodically turned off and on. Learning Objectives After going through this section session, you would be able to: • Know the basic definition of Laplace Transform. • Solve improper integral with an infinite limit. • Transform continuous functions like polynomials, exponentials, trigonometric (only cosine and sine ones) to Laplace forms • Find the inverse a Laplace transform back to polynomial or Laplace Transform. 4.1 DEFINITION: LAPLACE TRANSFORM Given a function f t defined for all t 0 , the Laplace transform of f is the function F of s defined as follows: L f t e st f t dt F s ……………….. (1.1) 0 for all values of s for which the improper integral converges. The improper integral over an infinite interval is defined as a limit of integrals over finite intervals, that is: m f t dt ………………………..(1.2) a f t dt mlim a 97 If the limit in (1.2) exists, then we say that improper integral converges. Otherwise, it diverges or fails to exist. Note that the integrand of the improper integral in (1.1) contains the parameter s in addition to the variable of integration t . Therefore, when the integral (1.1) converges, it converges not merely to a number, but to a function F of s . Example 4.1 Compute the Laplace transforms of the function (i) 1, (ii) t cost , for t 0 . eat , (iv) (iii) , Solution: m (i) 1 1 L1 1 e st dt lim e st for every s 0 0 m s 0 s (ii) 1 t Lt t e st dt lim e st e st dt 0 0 s m s 0 m m 1 1 lim 2 e st 2 for every s 0 m s 0 s e s a t at st s a t e e dt e dt 0 0 s a 0 (iii) L e at (iv) L cos at e st cos atdt 1 , sa sa s st e sin atdt a 0 0 m 1 1 lim e st sin at s e st sin atdt m a 0 0 a m 1 1 lim e st sin at s e st sin atdt m a 0 0 a s2 st e a2 a2 0 s cosatdt s a2 s2 a2 Lcosat s s2 s L cos at 2 2 L cos at 2 2 a a s a Find the following: L t 2 , L t 3 , L t 4 , and L t n 98 4.2 ELEMENTARY PROPERTIES OF LAPLACE TRANSFORM Theorem 4.1: (Linear Property) Let f and g be functions whose Laplace transforms exist, and let c1,c2 be any two real numbers, then: Lc1 f t c2 gt c1L f t c2 Lgt Prove as exercise. Example 4.2 Find the Laplace transforms of the given functions: f t 6e 5t e3t 5t 2 9 g t 4 cos 4t 9 sin 4t cos10t Solution: 1. F s 6 1 1 3! 1 5 31 9 s 5 s 3 s s F s 2. Gs 4 6 1 30 9 4 s 5 s 3 s s s s 4 Gs 2 2 9 4s s 16 2 4 s 4 2 36 s 16 2 2 2 s s 10 2 2 2s s 100 2 Example 4.3 1. Let f t cosh at . Find L f t . Solution: Lcosh at L 12 e at e at 12 L e at 12 L e at Lcosh at s s2 a2 , sa0 99 1 1 1 1 2 sa 2 sa 2. Let f t eiwt . Find L f t . Solution: L eiwt 1 s iw s iw s iw 2 s iw s iws iw s w2 s 2 w2 s 2 w2 By Euler formula: eiwt cos wt i sin wt L eiwt Lcoswt i sin wt Lcoswt iLsin wt Equating the real and imaginary parts of these two equations, we obtain the transforms of cosine and sine. Lcos wt s s w 2 2 , Lsin wt w s w2 2 . Theorem 4.2: (First Shifting Theorem) Replacement of s by s a in the function. If f t has the transformation F s (where s k ), then e at f t has the transform F s a (where s a k ). In formula: L e at f t F s a or, if we take the inverse on both sides e at f t L1F s a. Example 4.4 1. L e at coswt 2. L e at sin wt sa s a2 w2 w s a2 w2 100 Theorem 4.3: (Multiplying by t and t n ) If L f t F s then Ltf t F s . Because: de st L tf t tf t e st dt f t dt 0 0 ds For example; Lsin 2t Therefore, Since dn ds n d f t e st dt F s 0 ds 2 s 4 2 L t sin 2t d 2 4s . 2 ds s 4 s 2 4 2 e st 1n t ne st , we have: e f t dt 1 ds 1 Lt f t F n s 0 dn n st n e t f 0 st n n t dt n n n d n n F s 1 F n s Hence L t f t 1 n ds Theorem 4.4: Let f be a piecewise – continuous function on 0, exponential order. Define a function g by: 0t a 0 g t f t a a t where a positive number, then Lg e as L f t 101 Example 4.5 1. 0 0 t a 1 a t Let g t Then Lg e as 1 s 2. 0 t 0 cost t Let g t s Then e s s 1 2 Theorem 4.5: If L f t F s and if lim t 0 t 0 f t exists, then t 1 L f t F x dx t s Theorem 4.6: If L f t F s , then t 1 L f x dx F s 0 s Theorem 4.7: If f t is periodic with period , that is, f t f t , then L f t st 0 e f t dt 1 es Example 4.5 Further examples 1. et Find the Laplace transform of f t 3 102 t2 t2 Solution: 2 0 0 2 L f t e st f t dt e st et dt e st 3dt 1 e2 s1 3 2s L f t e s 0 s 1 s 2. Find the Laplace transform of the function graph below: 1 2 3 4 5 6 7 8 Solution: The function is 1 x 4 f t 1 x4 Therefore, 4 0 0 4 L f t e st f t dt e st 1dt e st 1dt 2e4 s 1 L f t s 0 s s 3. Find sin3t . L t Using Theorem 1.5, let can be written as: f t sin3t , then F s F x 3 x 9 2 103 3 , which also s 9 2 Therefore, sin3t 3 s L 3 dx tan 1 2 t s x 9 3 4. Find t L sinh 2 xdx 0 Solution: Taking f x sinh 2x , we have f t sinh 2t F s 2 , and then, from Theorem 1.6 that s2 4 t 1 2 2 L sinh 2 xdx 2 0 s s 4 s s 2 4 5. Find the Laplace for the square wave shown below Solution: Note that f t is periodic with 2 , and in the interval 0 t 2 it can be defined analytically by: 1 0 t 1 f t 1 0 t 2 From Theorem 1.7, we have: L f t 104 2 st 0 e f t dt 1 e2s Since 2 st 0 e f t dt e st 1 dt e st 1dt 1 2 0 1 2 1 2 s 1 e 2e s 1 e s 1 s s It follows that, 1 e F s s 1 e s 1 e 1 e s 1 e e s 1 2 1 e s 2 s s e s / 2 1 e s es / 2 s s 1 e 2 s s s e s / 2 e s / 2 1 s tanh s/2 s / 2 s 2 s e e Exercise4.1 Find the Laplace Transforms of the following functions 1. (i) f t sin t F s s 2 1 s 1 2 (ii) f t sinh t F s (iii) f t 2 x 2 3x 4 F s s43 s32 4s (iv) f t 2 tk k! . k 1 F s s 3 s 32 36 2. f t e3t cos6t 3. f t 3sinh 2t 3sin 2t 4. f t e3t cos6t e3t cos6t 5. f t t sin 2t 12s 2 16 F s s 43 6. f t t 2 F s f t tg t F s G s sG s 7. 2 3 105 3 5 4s 2 8. f t sin 3t cos 3t 9. f t 1 t 3 10. 1 0 t 1 f t 0 1 t 11. t 0 t 1 f t 0 1 t 12. Find the Laplace transform of the period function 0t t f t 2 t t 2 13. Find 1 s tanh 2 s 2 t1 L e4tt e4 x sin3xdx 0 x 2 s 4 2 1 3 s tan 1 2 3 s 4 s 2 9 s 4 106 4.3 INVERSE LAPLACE The inverse Laplace is given by: L1F s f t . Linearity of inverse Laplace transforms: L1aFs bGs aL1F s bL1Gs for any constant a, b . For example: 1/ s t , L L1 1 / s , Lt 1/ s 2 , L e at 1/s a , and Lcosat s / s 2 a 2 . Hence L11/ s 1 , L1 2 1 1/s a eat . Example 4.6 1. 2 3 1 1 L1 2 2L1 3L1 2 2 3t s s s s 2. s 5 1 1 1 s L1 2 5 L L 5e3t cos 2t 2 s 3 s 3 s 4 s 4 Example 4.7 Find the Laplace inverse of the function F s 2s 3 s 5s 4 Solution: By partial fraction: 2s 3 13 5 s 5s 4 9s 5 9s 4 2s 3 13 5 L1 L1 s 5s 4 9s 5 9s 4 13 1 1 5 1 1 13 5t 5 4t L L s 4 9 e 9 e 9 s 5 9 107 Exercise 4.2 Find the Laplace inverse of the following; 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. s s 6 cos 6t 2 s 5s 2 1 5 t sin t 2 2 s 1 s2 9 1 cosh3t sinh3t 3 s4 s 4s 8 e2t cos2t e2t sin2t 1 s 1 s 2 1 1 t 1 1 e cos t sin t 2 2 2 2 s 2 3s 1 s 2s 2 16 e as s s5 s 2 25 1 2s e e s s e s s s2 4 s2 s 23 108 4.4 CONVOLUTION Definition: Let f t and g t be piecewise – continuous functions on 0, of exponential order. The function called the convolution of f t and g t , denoted by f g , is defined by: f g t 0 f xg t xdx t Note: f g g f Theorem: Let f and g be piecewise continuous on 0, of exponential order. Then: L f g L f Lg or, equivalently, L1L f Lg f g Example 4.8 t 1 1. t 1 1 1 L 1 e 3t 1 e 3t x dx e 3t x 1 e 3t 0 3 ss 3 0 3 2. Using convolution, find the inverse f t of; F s 1 s 2 12 1 1 s2 1 s2 1 Solution: f t L1F s sin t sin t sin sin t d t 0 3. Let F s 1 t 1 t 1 1 cos td cos 2 t d t cos t sin t 2 0 2 0 2 2 1 s s a 2 . Find f t Solution: 1 L1 2 t s 1 at L1 e s a f t t e at xe at x dx e at xe ax dx t t 0 0 109 at e at 1 a2 1 4.5 SOLUTION OF LINEAR DIFFERENTIAL EQUATION The use of Laplace Transforms enables us to reduce the problem of finding a solution y to a linear differential equation with constant coefficients to the problems of solving a linear algebraic equation for L y and then determining y. L f t e st f t dt e st f t 0 s e st f t dt 0 0 L f t sL f t f 0 L f t s 2 L f sf 0 f 0 In general; L f n s n L f s n 1 f 0 s n 2 f 0 f n 1 0 Example 4.9 1. Let f t t 2 . Derive L f for L1 . Solution: Since f 0 0 , f 0 0 . f t 2 , and L2 2L1 2 / s , we obtain L f L2 2 / s s 2 L f Hence L t 2 2 / s 3 2. Let f t sin 2 t . Find L f . Solution: We have f 0 0 , Lsin 2t f t 2sin t cos t sin 2t 2 s2 4 sL f or ss 2 4 L sin 2 t 2 110 Example 4.10 1. Solve the differential equation (IVP) y0 2 y 3 y 1 Solution: Ly 3 y 1 Ly 3Ly L1 But Ly sLy y0 sLy 2 Hence sL y 2 3L y s 3L y 2 L y 1 s 1 s 2 1 s 3 ss 3 2 1 1 1 1 y L1 2L1 L s 3 s 3 ss 3 ss 3 1 1 5 1 y 2e 3t e 3t e 3t 3 3 3 3 2. Solve the differential equation (IVP) y 3 y 2 y e t y0 3, y0 4 Solution: Ly 3Ly 2Ly L e t ………..(I) but Ly s 2 Ly sy0 y0 Ly s 2 Ly 3s 4 …………….(II) and L y sLy y0 sLy 3 ………(III) Put (III) and (II) into (I) s 2 Ly 3s 4 3sLy 3 2Ly s 1 1 s 2 3s 2 L y 3s 5 L y 1 s 1 3s 5 1 s 2s 1 s 2s 1s 1 111 3s 5 1 y L1 s 2s 1 s 2s 1s 1 3s 5 1 1 y L1 L s 2s 1 s 2s 1s 1 by partial fraction, we have: 3s 5 1 1 1 1 1 1 1 1 1 y L1 L S 2 2 L s 1 6 L s 1 s 2s 1 3 y e2t 2et 13 e2t 12 et 16 et y 43 e2t 32 et 16 et Exercise 4.3 Work out the following (on convolution) 2. 1 sin t 1. 11 4. sin t cost 3. t et 5. eit e it Inverse Transforms by Convolution. Find f t by “the convolution theorem” 1. 6 ss 3 2. 1 s 2 s 1 3. 1 s s2 4 4. s2 s 2 2 2 5. e as ss 2 Solve the following differential equations using Laplace Transforms y0 0 , y0 0 1. y 2 y 5 y 5 2. 1 y 5 y 6 y 2 cos2t et y0 0 , y0 10 3. y y F t 4. y 2 y Asin t y0 1 , y0 0 5. y y F t 1 0 t 1 y0 0 , y0 0 ; where F t 0 1 t 4 0t 2 y0 0 , y0 0 where F t t 2 t 2 112 4.6 DIRAC’S DELTA FUNCTION Definition: Also f t t adt f a and t adt 1 . m m n f t t adt f a and n t adt , provided in each case, that n a m . In the Laplace transform form: L f t t a f ae as and L t a e as Note: t a is not a function in the ordinary sense. Example 4.11 1. 2 2t 2. 0 sin 5t t 2 dt f 2 sin 52 1 3. 0 2t 5 2 3 t 2dt f 2 2 22 3 11 3 2t 2 5t 6 t 1dt 7 5 2t e 2 t 3dt f 3 e 23 e6 4. 5. L t 3 e 3s 6. L2 t 3 f 3e 3s 2e 3s 7. L 2t 2 t 1 t 2 f 2e 2s 2 22 2 1 e 2s 7e 2s 8. L sin 3t t 2 sin 32 e 9. Lcosh2t t cosh20 e0 cosh0 1 11 1 s 2 113 s e 2 Exercise 4.4 Evaluate the following; 1. 2. 1 3 2t 4t 5 2 t 3dt 0 sinh 2t t dt 5 2t e 0 t 3dt 3. 4. 0 cos 5t t 2 dt Determine the following; 5. L2 t 1 6. L e 3t t 2 7. Lsin 3t t Solve the following differential equation; 8. y y t t 2 y0 0, y0 1 9. y 4 y 5 y t 1 y0 0, y0 3 10. y 2 y 3 y 8et t 12 y0 3, y0 5 114 UNIT 5 SYSTEM OF LINEAR DIFFERENTIAL EQUATIONS Introduction In the preceding course (MATH 251) we have discussed methods for solving Ordinary Differential Equations with constant coefficient that involves only one dependent variable. Many applications, however, require the use of two or more dependent variables, hence in this unit, we shall deal with more than one dependent variable and more than one equation. Learning Objectives After going through this unit, you would be able to: • put systems of differential equations in normal form. • reduce a higher order linear differential equation into first order system of differential equations. • put systems of equations in matrix form and solve. Session 1-1 Introduction and Motivation In real life situations quantities and their rate of change depend on more than one variable. For example, the rabbit population, though it may be represented by a single number, depends on the size of predator populations and the availability of food. In order to represent and study such complicated problems we need to use more than one dependent variable and more than one equation. Such problems lead naturally to a system of simultaneous Ordinary Differential Equations. Systems of differential equations are the tools to use. The differential equations are very much helpful in many areas of science. But most of interesting real life problems involves more than one unknown function. Therefore, the use of system of differential equations is very useful. 115 As a motivation let us consider an island with two types of species: Rabbits and Fox. Clearly one plays the role of predator while the other one the role of a prey. If we are interested to model the populations’ growths of both species, then we have to keep in mind that if, for example, the population of the Fox increases, then the Rabbit population will be affected. So the rate of change of the population of one type will depend on the actual population of the other type. For example, in the absence of the Rabbit population, the Fox population will decrease (and fast) to face a certain extinction, something that most of us would like to avoid. A model for this Predator-Prey problem was developed by Lotka (in 1925) and Volterra (in 1926) and is known as the Lotka-Volterra system dR aR RF dt dy bF RF dt where R t measures the Rabbit population, F t measures the Fox population, and all the involved constant a, b, , are positive numbers. Note that a and b are the growth rate of the prey, and the death rate of the predator. and are measures of the effect of the interaction between the Rabbits and The Fox. Can we check Iraq population with system of differential equation? Note that in the Lotka-Volterra system, the variable t is missing. This kind of system is called autonomous system and is written dy dt f x, y dy g x, y dt We will usually denote the independent variable by t and the dependent variables (the unknown functions of t ) by x1 , x2 , x3 ,, or x, y, z,. For instance, a system of two first order differential equations in the dependent variables x and y has the general form: 116 dx dy f t , x, y, , 0 dt dt ( ) dx dy g t , x, y, , 0 dt dt where the functions f and g are given. A solution of this system is a pair ( x(t ), y(t ) of functions of t that satisfy both equations identically over some real interval of values of t . 1-1.1 Types of Linear Systems Definition 1: The general linear systems of two first-order differential equations in two unknown functions x and y are of the form: a1 (t ) x a2 (t ) y a3 (t ) x a4 (t ) y f1 (t ) b1 (t ) x b2 (t ) y b3 (t ) x b4 (t ) y f 2 (t ) ............. (4.1) where the primes denote derivatives with respect to the independent variable t . i.e. x dx dt y dy dt Note: We shall be concerned with systems of this type that have constant coefficients and the number of equations Example 1.1 1. 3. 2 x 3 y 2 x y t t x 2 y 3x 5 y e 2 x y 2 x y e 2. t 4. x 3x y cos t 117 2 x 2 y x y t 1 x y x t 2 ( D 2) x ( D 1) y 1 ( D 1) x ( D 2) y t 1 We shall say that a solution of system (1.1) is an ordered pair real functions ( f , g ) such that x f (t ) , y g(t ) simultaneously satisfy both equations of the system (4.1) on some real interval a t b . Definition 2: The general linear systems of three first-order differential equations in three unknown functions x, y and z is of the form: a1 (t ) x a2 (t ) y a3 (t ) z a4 (t ) x a5 (t ) y a6 (t ) z f1 (t ) b1 (t ) x b2 (t ) y b3 (t ) z b4 (t ) x b5 (t ) y b6 (t ) z f 2 (t ) ..... (4.2) c1 (t ) x c2 (t ) y c3 (t ) z c4 (t ) x c5 (t ) y c6 (t ) z f 3 (t ) Example 1.2 x y 2 z 2 x 3 y z t 1. 2 x y 3z x 4 y 5z sin t x 2 y z 3x 2 y z cost 3. 2. x z 2 x 5 y z 1 t y 3x 5 y z tan t x y z 2 z e t D 1x1 x2 x3 0 2 x1 D 3x2 4 x3 0 4. 4 x1 x2 ( D 4) x3 0 dx1 3x1 2 x2 2 x3 dt dx2 x1 4 x2 x3 dt dx3 2 x1 4 x2 2 x3 dt We shall say that a solution of system (4.2) is an ordered triple of real functions f , g , h such that: x f (t ), y g (t ), z h(t ) simultaneously satisfy all three equations of the system (4.2) on some interval a t b . Systems of the form (4.1) and (4.2) contained only first derivative, and we now consider the basic linear system involving higher derivatives. This is the general linear system of two second-order differential equations in two unknown functions x and y , and is a system of the form: a1 (t ) x a2 (t ) y a3 (t ) x a4 (t ) y a5 (t ) x a6 (t ) y f1 (t ) ...... (4.3) b1 (t ) x b2 (t ) y b3 (t ) x b4 (t ) y b5 (t ) x b6 (t ) y f 2 (t ) An example is provided below with constant coefficient: 2 x 5 y 7 x 3 y 2 y 3t 1 3x 2 y 2 y 4 x y 0 118 For given fixed positive integers m and n , we could proceed in like manner, to exhibit other th general linear systems of n m - order differential equations in n unknown functions and give examples each of such type of system (I hope you can do that on your own). Instead we proceed to introduce the standard type of linear system. We now introduce standard type as a special case of the system (4.1) of two first-order differential equations in two unknown functions x and y . We consider the special type of linear system (4.1), which is of the form: x a11 (t ) x a12 (t ) y f1 (t ) .................. (4.4) y a21 (t ) x a22 (t ) f 2 (t ) This is called the normal form in the case of two linear differential equations in two unknown functions. The characteristic feature of such a system is apparent from the manner in which the derivatives appear in it. An example of such a system with variable coefficients is x t 2 x (t 1) y t 3 y tet x t 3 y e t , while one with constant coefficients is x 5x 7 y t 2 y 2 x 3 y 2t Note that example (4) of equation (4.2) is in the normal form. The normal form in the case of a linear system of three differential equations in three unknown functions x, y and z is: x a11 (t ) x a12 (t ) y a13 (t ) z f1 (t ) y a21 (t ) x a22 (t ) y a23 (t ) z f 2 (t ) z a31 (t ) x a32 (t ) y a33 (t ) z f 3 (t ) An example of such a system is the constant coefficient system: 119 x 3x 3 y z t y 2 x 4 y 5z t 2 z 4 x y 3z 2t 1 The normal form in the general case of a linear system of n differential equations in n unknown functions x1 , x2 ,, xn is: x1 a11 (t ) x1 a12 (t ) x2 a1n (t ) xn f1 (t ) x2 a21 (t ) x1 a22 (t ) x2 a2n (t ) xn f 2 (t ) ............. (4.5) xn an1 (t ) x1 an 2 (t ) x2 ann (t ) xn f n (t ) An important fundamental property of a normal linear system (4.5) is its relationship to a single n th – order linear differential equation in one unknown function. Specifically, consider the so-called normalized (meaning, the coefficient of the highest derivative is one) n th – order linear differential equation x ( n) a1 (t ) x ( n1) an1 (t ) x an (t ) x f (t ) ______ (4.6) in the one unknown function x . Let x1 x, x2 x x3 x ......................... (4.7) xn1 x ( n2) xn x ( n1) From (4.7), we get: x x1 , x x2 ,, x ( n1) xn 1 , x ( n) xn .......... (4.8) Then using both (4.7) and (4.8), the single n th – order equation (4.6) can be transformed into: 120 x1 x2 x2 x3 xn 1 xn xn an (t ) x1 an1 (t ) x2 a1 (t ) xn f (t ) .......... (4.9) which is a special case of the normal linear system (4.5) of n equations in n unknown functions. Thus we see that a single n th – order linear differential equation of the form (4.6) in one unknown function is indeed intimately related to a normal linear system (4.5) of n first – order differential equation in n unknown functions. Example 1.3 Reduced the following to normal form: (a) y(t ) a1 y(t ) a2 y(t ) a3 y(t ) f (t ) (b) d 2 x t dx e (t 1) x 0 dt dt 2 (c) d 2x sinh x 3x dt 2 Solution (a): y(t ) a1 y(t ) a2 y(t ) a3 y(t ) f (t ) ................ (α) Let y(t ) x0 (t ) y(t ) x0 (t ) x1 (t ) y(t ) x0(t ) x1 (t ) x2 (t ) y(t ) x0(t ) x1(t ) x2 (t ) then from (α), we have y(t ) f (t ) a1 y(t ) a2 y(t ) a3 y(t ) ................. (β) from above x0 (t ) x1 (t ) ............................................................. (γ) x1 (t ) x2 (t ) ............................................................. (δ) x2 (t ) f (t ) a1x2 (t ) a2 x1 (t ) a3 x0 (t ) ................ (ε) 121 system so equation (γ), (δ) and (ε) are the 1st – order linear differential equation equivalent to (α). These three (3) equations can be solved to obtain the solution to (α). Solution (b): x x0 Let dx dt dx0 dt 2 x1 d x dt 2 2 d x0 dt 2 dx1 dt e t dx dt t 1 x The normal form is: dx0 dt dx1 dt x1 e x1 (t 1) x0 t Exercise 1.2 In each of exercises 1 – 4, transform the single linear differential equation of the form (1.6) into a system of first – order differential equations of the form (1.9). 1. dx 2 dx 3 2x t 2 2 dt dt 2. d 3x d 2 x dx 2 2 x e 3t 3 2 dt dt dt 3. d 3x d 2x dx t 2t 3 5t 4 0 3 2 dt dt dt 4. d 4 x 2 d 2 x dx t 2tx cost dt 4 dt 2 dt 2. dx dy 2 2x y dt dt Example 1.5 Reduce the following to normal form; 1. 2 dx dy 2 x 5 y 2t dt dt dx dy 2 x 3t 0 dt dt 2 Solution: 2 dx dy 2 x 5 y 2t _________(1) dt dt 122 dx dy x y dt dt dx dy 2 x 3t ____________(2) dt dt Solving for dy dx by the method of elimination (i.e. eliminating first) then (1) 2 (2) dt dt 5 Also solving for dx dx 1 3x 5 y t 3x y t dt dt 5 dy then (1) (2) 2 dt 5 dx dy 4 8 4 x 5 y 8t x y t dt dt 5 5 Therefore we have the normal linear system as: dx 3 1 x y t dt 5 5 dy 4 8 x y t dt 5 5 Exercise 1.3 Find the first order system of the following differential equations 1. d2y 4y 0 dt 2 2. d3y d2y dy 9 3 12 2 4 cost dt dt dt 3. d 2x dx m 2 c kx f (t ) , where f (t ) is an external force. dt dt 4. d 3x d 2x dx 3 3 2 2 12 8x e t sin t dt dt dt 5. d 4x d 2x 5 2 4 x e t te 2t 4 dt dt d5x d 3x d 2x 2 2 3t 2 1 6. 5 3 2 dt dt dt 123 Reduce each of these to the normal form. 1. dy dx 2 et dt dt dy dx x y sin t dt dt 2. dx dy 2 x et dt dt dx dy 3 2 x 4 y sinh t dt dt 3. dy dx x 2 y sin t dt dt dy dx 4e t x dt dt 4. d 2x dx dy 6 3 0 2 dt dt dt 2 d y dx x y 4t dt 2 dt 1-1.2 Solution to a Linear System of Differential Equations The solution to the system: dx1 f1 t , x1 , x2 , x3 , xn dt dx2 f 2 t , x1 , x2 , x3 , xn dt ....................... (1.10) dxn f n t , x1 , x2 , x3 , xn dt x1 (t ) x (t ) 2 on open real interval a t b for which the system is is the parametric vector x xn (t ) true and the vector x must be continuous on the open real interval. Session 2-1 Methods of Solving SLDE The most elementary approach to Systems of Linear Differential Equations (SLDE) with constant coefficients involves the elimination of dependent variables by appropriately combining pairs of equations. The main methods we will consider in solving such systems are: A) The differential operator method, B) The matrix method. 124 The differential operator method for linear differential system is quite similar to the solution of linear algebraic systems by elimination of variables until only one remains. It is most convenient in the case of manageably small systems: those consisting of no more than four equations. For large systems of differential equations, the matrix method is preferable. 2-1.1 The Differential Operator Method Recall that D d d2 dn D 2 2 and also D n n dt dt dt System of 1st Order Equations: The differential operator method for linear differential system is quite similar to the solution of linear algebraic systems by elimination of variables until only one remains. It is most convenient in the case of manageably small systems: those consisting of no more than four equations. For large systems of differential equations, the matrix method is preferable. Example 1.6 Solve the system dx x y dx dy dt dy x y 4x y 4 x y dt Note: this system is called Autonomous system since time is not explicitly expressed in the equation. Solution: We use matrix/determinant techniques. We rewrite the problem as follows: D 1x y 0 4 x D 1 y 0 D 1 1 x 0 y 0 4 D 1 Using Cramer’s Rule: 125 D 1 1 0 1 x 4 D 1 0 D 1 D 1 2 4 x 0 D 2 2D 3 x 0 x 0 D 3,1 (Just like finding the roots of a characteristic equation in linear differential equation we did last semester) x Ae3t Be t Then go back to the system to pick one equation to determine y . y dx x 3 Ae3t Bet Ae3t Bet 2 Ae3t 2Bet dt Then the solution is written as: x(t ) Ae 3t Be t y(t ) 2 Ae 3t 2Be t Example 1.7 1. Solve the system dx dy , when x(0) 1 , y(0) 2 3x y x y Solution: dx 3x y D 3x y 0 1 x 0 D 3 dt 1 D 1 y 0 dy x y x D 1y 0 dt Using Cramer’s Rule: D3 1 0 1 D 3 D 1 1 x 0 x 1 D 1 0 D 1 D D2 D 3D 4 x 0 2 4D 4 x 0 D 22 x 0 but x 0 x At Be 2t _________(*) Solving for y , we have: y 3x dx dt y 3e 2t At B A 2At Be 2t 126 y 3e 2t At B 2 At 2B Ae 2t y Ate 2t B Ae 2t _____ (**) with the initial conditions: (*) B 1 (**) 2 B A A 1 x 1 t e 2t y te2t 2e 2t 2. Solve the system: D 1 x Dy 2t 1 2D 1 x 2Dy t Solution: D 1 x Dy 2t 1........() 2D 1 x 2Dy t...........() Subtracting twice (I) from (II), we have: 3x 3t 2 . Substituting x t 2 / 3 in (I), we obtain: Dy 2t 1 D 1 x t 43 y 12 t 2 43 t c The complete solution: x t 23 , y 12 t 2 43 t c . Exercise 4.4 1. dx y dt dy x dt 4. dx 2 x y dt dy x y dt 2. dx y dt dy x dt 5. dx 3x 2 y t dt non-homogeneous dy 2x 3 y 3 dt 127 3. dx 6x 8 y dt dy 6x 2 y dt 6. dx 6x 8 y dt dy 6 x 2 y e t dt 8. dx1 3x1 x2 dt dx2 3x1 2 x2 dt dx3 x1 x2 x3 dt 10. 7. dx 3x 2 y 2 z dt dy x 4y z dt dz 2 x 4 y z dt 9. dx1 x1 x2 x3 e t dt dx2 x1 x2 t dt dx3 x1 x3 1 dt Find the solution of the system of differential equation Ex. 1 such that x0 1 , y0 2 . 11. Find the solution of the system of differential equation is Ex.5 such that x0 3 , y0 1 . 12. Find the solution of the system of differential equations in Ex.7 such that x0 2 , y0 1 , z0 0 . 13. Find the solution of the system of differential equations in Ex.9 such that x1 0 1 , x2 0 1 , x3 0 0 . 128 2-1.2 Reduction of nth – Order Equation to a System of 1st Order Equations The number of integration constants in an n th order differential equation, in the same vein, is the same as the number of integration constant in a system of n 1st order differential equation. Example 1.8 1. Convert y 3y 2 y et into a system of equations and solve completely. Solution: y(t ) x0 (t ) Let y(t ) x0 (t ) x1 (t ) y(t ) x1 (t ) The system becomes: x0 (t ) x1 (t ) non – homogeneous example x1 (t ) et 3x1 (t ) 2 x0 (t ) Then the solution of the system: D 1 x0 0 t 2 D 3 x1 e Using Cramer’s Rule: D 1 0 x1 t 2 D3 e 1 D3 DD 3 2x1 et D2 3D 2 x1 et D 2D 1x1 et Let D 2x1 z D 1z et z z et e t 1 t e D 1 z tet k1et e t dt Then D 2x1 tet k1et x1 e 2t tet k1e t e 2t dt 129 x1 e 2t te t k1e t dt x1 e 2t tet e t k1e t k2 x1 et 1 t k1e t k2 e 2t Since we are solving for y(t ) , then we have y Ae t Be 2t et 1 t Exercise 1.4 y 7 y 12y t 1. D D 3. 5. 7. D 2 2. D 32 y te3t DD 1y 5 2 5D 6 y e 2t 4. 2 2D y e x cos x 6. 1 y tan t 2 , 2 t 2 8. D D 2 2 Dy 2 D y x t 2-1.3 The Matrix Method Although the simple differential operator method suffice for the solution of small linear systems containing only two or three equations with constant coefficients, the general properties of linear systems – as well as solution methods suitable for large systems – are most easily and concisely described in matrix notation. We discuss here the general system of n first order linear equations being the form: (note equation 4.5) dx1 a11 (t ) x1 a12 (t ) x2 a1n (t ) xn f1 (t ) dt dx2 a21 (t ) x1 a22 (t ) x2 a2 n (t ) xn f 2 (t ) ______ (4.5) dt dxn an1 (t ) x1 an 2 (t ) x2 dt ann (t ) xn f n (t ) where f1 t , f 2 t , f 3 t ,, f n t are continuous on a real interval a t b . 130 The Matrix Notation form of (4.5) is given as: dx1 dt a11 a12 dx2 dt a21 a22 dx3 a a32 31 dt a dxn n1 an 2 dt This can simply be written as: a13 a23 a33 an3 a1n f1 t a2n f 2 t a3n f 3 t ann f n t dx Ax F(t ) …………............... .(I) dt which is a NON-HOMOGENEOUS FORM. where F(t ) is the non-homogeneous term or the force term. If F(t ) 0 , that is f1 (t ) f 2 (t ) f 3 (t ) f n (t ) 0 for the real interval a t b then we have dx Ax ………………............... (II) dt Equation (II) is known as the HOMOGENEOUS FORM. Examples of non – homogeneous systems of linear differential equation 1. dx x y e t dt dy x y 6sin t dx The matrix notation form is given as: dx dx dt 1 1 x et 1 1 x dx dt , x and , A where dt dy dy 1 1 y 6sin t 1 1 y dt dt e t F(t ) 6 sin t 131 2. dx1 3x1 x2 2 x3 t 2 dt dx2 4 x1 2 x2 x3 e t , dt dx3 2 x 2 5 x3 sin t dt The matrix notation form is given as: dx1 dx1 2 dt 1 2 x1 t 1 2 3 x1 dt 3 dx dx2 t dx2 4 2 1 x2 e where , A 4 2 1 , x x2 dt dt dt 0 2 x 5 x3 sin t dx3 0 2 dx3 5 3 dt dt t2 and F(t ) e t sin t Note that dx Ax F(t ) can be simply be written as x Ax Ft . dt Examples of Homogeneous systems of linear differential equation 1. dx 5x 3 y dt dy 2 x 4 y dt dx dx dx dt dt 5 3 x The matrix notation form is given as : where , dy 2 4 y dt dy dt dt 5 3 x and x A 4 2 y 2. dx1 x1 x2 x3 dt dx2 2 x1 4 x3 dt dx3 2 x1 3x3 dt 132 The matrix notation form is given as: dx1 dx1 1 1 x1 1 1 1 x1 dt 1 dt dx dx2 dx2 2 0 4 x2 where , A 2 0 4 and x x2 dt dt dt 0 2 x 3 x3 dx3 0 2 dx3 3 3 dt dt Note that dx Ax can simply be written as x Ax dt Definition Let A and F be defined on an open interval I. A vector function u is a solution of equation (I) on I if: u(t ) Au(t ) F(t ) Example 1.10 Consider the differential equation 5 3 x -------------- (1i) x 4 2 3e 7t is a solution of equation (1i) on The vector function: u(t ) 7t 2e 21e 7t 5 3 3e 7t 21e 7t (, ) since u(t ) and 7t 4 2e 7t 14e 7t 2 14e e 2t 2e 2t Similarly, v(t ) 2t is a solution of equation (1) on (, ) since: v(t ) 2t and e 2e 5 3 e 2t 2e 2t 2t 2t 2 4 e 2e 133 Example 1.11 Consider the equation: 1 1 1 x 2 3 4 x ----------- (1ii) 4 1 4 e 3t e 2t 3t The vector functions: u(t ) 7e and v(t ) 2e 2t are solutions on (, ) of equation 11e 3t e 2t 3e 3t 2e 2t (1ii) since: u(t ) 21e 3t v(t ) 4e 2t and 33e 3t 2e 2t 3t 3t 1 1 1 e 3e 3t 3t 2 3 4 7e 21e 4 1 4 11e 3t 33e 3t 2t 2t 1 1 1 e 2e 2t 2t 2 3 4 2e 4e 4 1 4 e 2t 2e 2t et It is now left for you to verify that : w(t) e t is also a solution on (, ) of equation (1ii) et Theorem 1.1 If u and v are solutions on an open interval I of x Ax and c1 ,c2 are any two constants, then w(t ) c1u(t ) c2 v(t ) is also a solution on I . Example 1.12 3e 7t e 2t In example 1.10 we found that u(t ) and v(t ) 2t are solutions 7t 2 e e 5 3 x of x 2 4 134 3e 7t e 2t 3c1e 7t c2 e 2t is also a solution for any choice c By Theorem 1.1, c1 2 2t 7t 2t 7t 2e e 2c1e c2 e of the constants c1 and c 2 . The vector – valued functions x1 , x 2 ,x n are linearly dependent on the interval I provided that there exist constants c1 , c2 ,, cn not all zero such that: c1x1 (t ) c2 x 2 (t ) cn x n (t ) 0 for all t in I . Otherwise they are linearly independent. Equivalently, they are linearly independent provided that no one of them is a linear combination of the others. 135 Exercise 1.5 a. Write the given system in the form: x(t ) A F(t ) 1. x 3x 2 y y 2x y 2. x x 2 y cost y x y sin t 6. 3. x 2 x 4 y 3e t y 5x y t 2 x 3x 4 y z t 4. y x 3z t 2 z 6 y 7z t 3 x 4 x y z 5. y 2 x y z z x 3y z State whether homogeneous or non – homogeneous. b. Verify that the given vectors are solutions to the given system, and then use the Wronskian to show that they are linearly independent. Finally write the general solution of the system. 1. e 3t 2e 2t 3 2 x ; x1 3t , x 2 t x 3 4 3e e 2. 3 1 1 1 x ; x1 e 2t , x 2 e 2t x 5 3 1 5 3. 4 1 1 2 x ; x1 e 3t , x 2 e 2t x 2 1 1 1 4. 0 1 1 1 1 0 2t t t x 1 0 1x ; x1 e 1 , x 2 e 0 , x 3 e 1 1 1 0 1 1 1 5. 0 3 2 2 2 2 t 3t 5t x 1 3 2 x ; x 1 e 2 , x 2 e 0 , x 3 e 2 0 1 1 1 1 3 136 Theorem 1.2: Principle of Superposition Let x1 , x 2 ,x n be n solutions of homogeneous linear equation (II) i.e.: dx Ax on the dt open interval (, ) . If c1 , c2 ,, cn are constants, then the linear combination, x c1x1 (t ) c2 x 2 (t ) cn x n (t ) is also a solution of (II). Prove as exercise. 2-1.4 The Eigenvalue Method of Homogeneous Linear Systems In this section we will present a method for finding a general solution of the differential equation: dx Ax ………. (III) dt where A is an n n matrix. Recall that ordinary differential equation: Dy ay or D 2 aD b y 0 , has solutions of the form: y(t ) ce rt , where c is an arbitrary constant, r is the root(s) of the characteristic. We will show that equation (III) has analogous function for solutions in system of differential equation. That is, we will show that equation (III) has solutions of the form: x(t ) vet where is a constant (possibly a complex number) and v is a constant. To begin, we note that if: v then: e t1 e t1 1 t t d t d e 2 e 2 t 2 e v e e t v dt dt e t e t n n n Thus x(t ) vet is a solution of x Ax if and only if e t v et Av or, equivalently, if and only if Av v 2-1.5 Definition: Eigenvalue and Eigenvector The number (either zero or non – zero) is called an eigenvalue of n n matrix provided: det(A I) A I 0 137 An eigenvector associated with the eigenvalue is a non-zero vector v such that Av v , so that: A v 0 Lemma: If x(t ) v k e k t is a solution to (III) for the k th eigenvalue of A of order n n then: x(t ) c1 v1e 1t c2 v 2 e 2t cn v n e nt is also a solution, where ci is a constant for i 1,2,, n In outline, the eigenvalue method for solving system x Ax goes as follows: (I) Solve the characteristic equation det(A I) A I 0 for the eigenvalues 1 , 2 ,, n of matrix A . (II) We attempt to find n linearly independent eigenvector v1 , v 2 ,, v n (so that v k 0 for each k ) associated with these eigenvalues. That is A Iv k 0 for each k . This will not always be possible, but when it is, we get n linearly independent solutions: x1 (t ) v1e t , x 2 (t ) v 2 e t ,, x n (t ) v n e t (III) The general solution is then a linear combination of these n solutions : x(t ) c1 v1e t c2 v 2 e t cn v n e t We will discuss the various cases that can occur separately, depending upon whether the eigenvalues are distinct or repeated, real or complex. Case 1: Distinct Real Eigenvalues If the eigenvalues 1 , 2 ,, n are real and distinct, then we substitute each of them in turn in A I 0 and solve for the associated eigenvectors Example 4.13 Find a general solution of the system x1 4 x1 2 x2 x2 3x1 x2 138 v1 , v 2 ,, v n . Solution: 4 2 x The matrix form of the system is: x 3 1 The characteristic equation of the coefficient matrix is: 4 2 0 3 1 4 1 6 0 2 3 10 0 2 5 0 So we have the distinct real eigenvalues: 1 2 and 2 5 For the coefficient matrix A the eigenvector equation A I 0 takes the form: 2 1 0 4 ………..(δ) 1 2 0 3 When we substitute the first eigenvalue 1 2 , we find that the two scalar equations that follow are: 61 2 2 0 and 31 2 0 . These equations are equivalent i.e. they are redundant, so only one is needed; we can choose 1 arbitrary (but non-zero) and solve for 2 . The simplest choice is 1 1 , which yields 1 2 3 , and thus: v1 is an eigenvector associated with 1 2 (as is any zero 3 constant multiple of 1 ). Remark: If, instead of the “simplest” choice 1 1 , 2 3 , we had made another choice, 1 c , 2 3c , we would have obtained the eigenvector c 1 c v1 3c 3 Because this is a constant multiple of our previous result, any choice we make leads to the 1 same solution: x1 (t ) e 2t 3 139 Next we substitute in (δ) the second eigenvalue 2 5 and get the equivalent scalar equations 1 2 2 0 and 31 6 2 0 . 2 With 2 1 we obtain 1 2 , so: v 2 is an eigenvector associated with 2 5 . A 1 different choice 1 2c , 2 c would merely give a (constant) multiple of v 2 . 1 These two eigenvalues and associated eigenvectors yield the two solutions x1 e 2t 3 2 and x 2 e 5t . 1 e 2t They are linearly independent because their Wronskian 3e 2t 2e 5t 7e 3t 0 , since 5t e e 3t 0 is nonzero. Hence a general solution of the system is: 1 2 x c1x1 c2 x 2 c1 e 2t c2 e 5t ; 3 1 in scalar form: x1 c1e 2t 2c2 e5t , x2 3c1e 2t c2 e5t . Case 2: Repeated Real Eigenvalues Suppose that the n n matrix A has m n distinct eigenvalues 1 , 2 ,, n . Then at least one of these eigenvalues is a repeated root of the characteristic equation: A I 0 . An eigenvalue is of multiplicity k if it is a k fold root. Let 1 and 2 be eigenvalues of the system matrix such that 1 2 then the system has two linearly independent solutions of the form: x c1x1 c2 x 2 , where x1 ve t and x 2 vtet wet . And we use the equation A I w v for the second part that is x 2 . Example 1.15 Find the general solution of the system 140 x1 3x1 x2 x2 x1 x2 Solution: The characteristic equation is: 3 1 22 0 1 2 2 1 1 1 1 0 3 1 2 0 1 The eigenvector will be solution of: For 0 1 1 2t 2, 1 2 e 2t then v1 x1 v1e 1 1 1 2 0 With this eigenvector v1 and eigenvalue 2 we use the equation: A I w v1 1 1 1 1 1 2 1 1 1 2 1 Let 1 0 then 2 1 Therefore x 2 v1te 2t 1 0 we 2t te 2t e 2t 1 1 1 Hence x c1x1 c2 x 2 c1 e 1 2t t 2t e . c2 t 1 Use WRONSKIAN to determine whether their linearly independent. Case 3: Complex Eigenvalues To find the general solution of a two – dimensional system x Ax , where the eigenvalues of A are nonreal (i.e. complex conjugates), i : I. Find a complex eigenvector v v1 iv 2 corresponding to an eigenvalue . II. Construct a complex – valued solution of the system as: t xe cos t i sin t v , i . III. Separate the solution into a real part x1 and an imaginary x2 . IV. The general solution of the system is x c1x1 c2x2 141 Example 1.14 1. Find a general solution of the system x1 x1 4 x2 x2 2 x1 3x2 Solution: The characteristic equation of the coefficient matrix is given as: 1 4 1 3 8 2 2 5 0 2 3 1 2i , Note: the complex roots are conjugates. Next we choose 1 2i and find an eigenvector. Thus, The eigenvectors will be solutions of: 4 1 1 2i v1 0 2 3 1 2 i v2 0 4 v1 0 2 2i 2 2i v2 0 2 2 2i v1 4v2 0 2v1 2 2i v2 0 As before, the equation corresponding to the tow row are redundant, so only one is needed. Using the top row, we obtain the equation: 2 2i v1 4v2 0 or v2 1 i / 2v1 we may avoid fractions by choosing v1 2 ; then v2 1 i . Hence, we have the eigenvector: 2 v 1 i t A complex – valued solution is constructed from the formula x e v and Euler’s formula: 2 12i t e 1 i x 2cos t 2i sin 2t 2 t t e cos 2t i sin 2t e 1 i 1 i cos 2t 1 i sin 2t The real part of this solution is: x1 e t 2cos 2t cos 2t sin 2t and the imaginary part is: 2sin 2t t x2 e cos 2t sin 2t 142 Both of these real-value functions are solutions of the system, and with them we get a general solution: x x1 x 2 2cos 2t 2sin 2t t x e c1 c2 cos 2t sin 2t cos 2t sin 2t Case 4: Systems with Zero as an Eigenvalue We discussed the case of system with two distinct real eigenvalues, repeated (nonzero) eigenvalue, and complex eigenvalues. But we did not discuss the case when one of the eigenvalues is zero. Example 1.15: Find the general solution to x1 2 x1 x2 x2 2 x1 x2 Solution: The procedure for solving a system with zero eigenvalue is the same as system with distinct real eigenvalue. 2 1 0 2 1 The characteristic polynomial of this system is: 2 1 0 0 , which reduces to 2 2 3 0 . The eigenvalues are 1 0 and 2 3 . Let us find the associated eigenvectors; For 1 0 , A I1 v1 0 2 0 1 v1 0 2 1 0 v 0 2 2v1 v2 0 2v1 v2 0 The equation A I1 v1 0 translates into: 1 2 The two equations are the same. So we have 2v1 v2 . Hence an eigenvector is: v1 143 v1 For 2 3 , set: v 2 v2 2 3 1 v1 0 v 0 or 2 1 3 2 The equation A 3I v 2 0 translates into: v1 v2 0 2v1 v2 3v1 2v1 v2 3v2 2v1 2v2 0 The two equations are the same (as v1 v2 0 ). So we have v2 v1 . Hence an 1 eigenvector is: v 2 . 1 0t 1 3t 1 1 3t 1 Therefore the general solution is: x c1e c2 e c1 c2 e 2 1 2 1 Exercise 1.6 a. Apply the eigenvalue method of this section to find a general solution of the given system. If initial values are given, find the corresponding particular solution. 1. x1 x1 2 x2 x2 2 x1 x2 4. x1 2 x1 3x2 x2 2 x1 x2 2. x1 3x1 4 x2 ; x (0) x2 (0) 1 x2 3x1 2 x2 1 5. x1 4 x1 x2 x2 6 x1 x2 3. x1 6 x1 7 x2 x2 x1 2 x2 6. x1 9 x1 5 x2 ; x2 6 x1 2 x2 b. Find general solutions of the systems in these exercises. 1. x1 5x1 6 x3 x2 2 x1 x2 2 x3 x3 4 x1 2 x2 4 x3 x1 3x1 2 x2 2 x3 3. x2 5x1 4 x2 2 x3 x3 5x1 5x2 3x3 144 2. x1 3x1 x2 x3 x2 5x1 3x2 x3 x3 5x1 5x2 3x3 4. x1 2 x1 x2 x3 x2 4 x1 3x2 x3 x3 4 x1 4 x2 2 x3 x1 2 x1 3x2 3x3 x2 3x3 has the c. Show that the coefficient matrix of the system x2 x3 2 x3 eigenvalue 2 of multiplicity 2. Find two linearly independent eigenvectors associated with 2 , and then find a general solution of the system. d. i) Show that the coefficient matrix of the system x1 3x1 x2 x2 x1 x3 x3 x1 2 x2 3x 3 has the single eigenvalue 2 of multiplicity 3 , but with only one linearly independent eigenvector u associated with it. Then one solution is x1 (t ) ue 2t . ii) Find a second solution of the form x 2 (t ) ute2t ve 2t where v is a nontrivial solution of A 2I v u . iii) Find a third solution of the form x 3 (t ) 12 ut 2 e 2t vte 2t we 2t where w is a nontrivial solution of A 2I w v . e. Solve for the system x y 2 x 4 y 2 x 3 y 2 x 6 y Hint: the system above is not in the normal form, so reduce it to normal form first before solving it. 145 2-1.6 Non–Homogeneous Linear Systems Here we discuss two ways (methods) of solving non – homogeneous linear systems. The methods are; I. Undetermined coefficients II. Variation of parameters These are analogous to finding a particular solution of a single non-homogeneous n th order linear differential equation. Given the non-homogeneous first order linear system: x(t ) A(t )x F(t ) …………… (1) The general solution of (1) has the form: x(t ) x c (t ) x p (t ) …………….(2) where x c denotes a general solution of the associated homogeneous system: x(t ) A(t )x and x p is a single particular solution of the equation (1) I. Undetermined Coefficient The method is essentially the same for a system as for single differential equations – we make an intelligent guess as to the general form of a particular solution x p and then attempt to determine the coefficients in x p by substitution in x(t ) Ax F(t ) . Example 1.16 Find a general solution of the system x 4 x 2 y 8t ………...... (3) y 3x y 2t 3 Solution: We found the general solution x 1 2 x c c c1 e 2t c2 e 5t ……….... (4) 3 1 yc of the associated homogeneous system. Because there is no duplication between the terms x c and the non – homogeneous terms, we assume a trial solution of the form: 146 xp a b x p at b 1 t 1 ……....... (5) a2 b2 yp Upon substitution of (5) into (3) we get: a1 4 2 a1t b1 8t a2 3 1 a2 t b2 2t 3 4a 2a 2 8 4b1 2b2 t 1 3a1 a2 2 3b1 b2 3 When we equate the coefficients of t and constant terms, we get the equations: 4a1 2a2 8, 3a1 a2 2, 4b1 2b2 a1 , 3b1 b2 a2 3 hence a1 52 , a2 165 , b1 2 25 and b2 1 25 . 52 t 252 and the general solution of the system is given Thus our particular solution is: x p 16 1 t 25 5 as: x c1e 2t 2c2 e 5t 52 t 252 y 3c1e 2t c2 e 5t 165 t 251 Example 1.17 Find a general solution of the system x 4 x 2 y, y 3x y e 2t …………….(6) Solution: The complimentary solution is given as: x 1 2 x c c c1 e 2t c2 e 5t …….. (7) 3 1 yc Because of duplication of the term e 2t in non-homogeneous system, we assume a trial xp a1t b1 2t e ……(8) solution of the form: x p ate 2t be 2t y a t b p 2 2 (rather than ate2t alone). Then xp (a 2b)e 2t 2ate 2t ……. (9) 147 Substitution of (8) and (9) into (6), and cancellation of e 2t throughout, yields: a1 2b1 2a1t 4 2 a1t b1 0 a2 2b2 2a2 t 3 1 a2 t b2 1 If we equate the coefficients of t and the constant terms, we get the equations: 6a1 2a2 0, 3a1 a2 0, 6b1 2b2 a1 , 3b1 b2 a2 1 The first two equations imply that a2 3a1 . In order for the last two equations to be consistent, we must have a1 2(a2 1) 2(3a1 1) 6a1 2 . It follows that a1 72 , and so a2 76 and b2 3b1 17 Hence the general solution of the system is given by: x c1e 2t 2c2 e 5t 72 te 2t y 3c1e 2t c2 e 5t 76 te 2t 17 e 2t Fundamental Matrices Suppose that x1 (t ), x 2 (t ),, x n (t ) are n linearly independent solutions on some open interval of the homogeneous system x Ax of n linear equations. Then the n n matrix x11 (t ) x12 (t ) x1n (t ) x21 (t ) x22 (t ) x2 n (t ) Φ(t ) x (t ) x (t ) x (t ) n2 nn n1 having as column vectors the solution vectors x1 , x 2 ,, x n is called a fundamental matrix for the system. Because its column vectors are linearly independent, it follows that the matrix Φ(t ) is non-singular and therefore has an inverse matrix Φ1 (t ) . In terms of the fundamental matrix Φ(t ) , the general solution: x c1x1 (t ) c2 x2 (t ) form: x(t ) Φ(t )c ………….(α) where c c1 , c2 ,..., cn 148 cn xn (t ) can be written in the In order that the solution x(t ) satisfy the initial condition: x(t 0 ) b where b1 , b2 ,bn are given, it therefore will suffice for the coefficient vector c in x(t ) Φ(t )c to satisfy 1 Φ(t0 )c b ; that is: c Φ (t 0 )b …………….(β) When we combine (α) and (β), it becomes clear that the solution of the initial value problem x Ax , x(t 0 ) b is given by: x(t ) Φ(t )Φ 1 (t 0 )b ……………(γ) Example 1.18 Find the fundamental matrix for the system: x 4 x 2 y and use it to find the solution of the y 3x y system which satisfies the initial conditions x(0) 1, y(0) 1 . Solution: e 2t 2e 5t 5t yield the fundamental x ( t ) The linearly independent solutions: x1 (t ) and 2t 3e e e 2t matrix: Φ(t ) 2t 3e 2e 5t e 5t 1 2 and the inverse matrix is given as: Φ 1 (0) Then Φ(0) 3 1 1 7 1 2 3 1 Hence the formula (γ) gives the solution: e 2t x(t ) 2t 3e 2e 5t 1 1 2 1 7 1 1 e 5t 2 1 7 3e 2t 4e 5t 2t 5t 9 e 2 e Thus the solution of the original initial value problem is given by x 73 e2t 74 e5t y 97 e2t 72 e5t 149 II. Variation of Parameters Theorem If Φ(t ) is the fundamental matrix of dx dx Ax . Let x p be a solution to Ax F(t ) , then dt dt we have the following; x p (t ) Φ(t ) Φ 1 (t )F(t )dt x(t ) Φ(t )c Φ(t ) Φ 1 (t )F(t )dt ………….. (ζ) Then also with definite integration we have x p Φ(t ) Φ(s)F(s)ds t t0 If we add this particular solution and the homogeneous solution, we get the solution: x(t ) Φ(t )Φ 1 (t 0 )b Φ(t ) Φ 1 (s)F(s)ds ……….(ε) t t0 of the initial value problem x(t ) A(t )x F(t ), x(t0 ) b . Example 1.19 4 2 15 2t 1 x te , x(0) Solve the initial problem x 3 1 4 1 Solution: If we were to use the method of undetermined coefficients, our trial solution would be of the form x p (t ) at 2 e 2t bte 2t ce 2t , and we would have six scalar coefficients to determine. Here it will be simpler to use the method of variation of parameters. e 2t We already know the fundamental matrix: Φ(t ) 2t 3e 2e 5t of the associated e 5t homogeneous system. The determinant , Φ(t ) 7e3t , so the inverse matrix: e 2t Φ 1 (t ) 17 e 3t 2t 3e 2e 5t e 5t From the theorem gives the particular solution: x p Φ(t ) e 3 s Φ (t ) e 3 s t 1 0 7 t 1 0 7 e 5s 2 s 3e 2e 5s 15se2 s ds e 2 s 4se2 s 7se3s ds 4 s 49se 150 t s ds Φ (t ) 0 7 se 7 s s t 12 s 2 Φ (t ) 7 s 1 7 s se 7 e s 0 e 2t 2t 3e 12 t 2 2e 5t e 5t te 7t 17 e 7t 17 Therefore: 4 28t 7t 2 1 5t 2 7 e x P (t ) 141 e 2t 2 2 14 t 21 t 1 This is a particular solution such that x p (0) 0 . 3e 2t 4e 5t of the We already know (from previous examples) the solution: x c (t ) 17 2t 5t 9e 2e 1 associated homogeneous systems such that: x c (0) 1 Hence the solution of the initial value problem is given by: 16 14t 7t e x 141 10 28t 7t 2 e 2t 72 e 5t y 141 2 2t 17 e 5t Summary: The Method of Solving Non-Homogeneous systems dx Ax . dt 1. Solve for the homogeneous part 2. Form your fundamental matrix Φ(t ) 3. Find Φ1 (s) 4. Evaluate 5. Find x p Φ(t ) Φ(s)F(s)ds . 6. Your final step x(t ) Φ(t )Φ 1 (t 0 )b Φ(t ) Φ 1 (s)F(s)ds . Φ t t0 1 (s)F(s)ds t t0 t t0 151 Exercise 1.7 a. Apply the method of undetermined coefficients to find a particular solution of each of the system below. If initial conditions are given, find the particular solution that satisfies these conditions. 1. 2. 3. x x 2 y 3 y 2x y 2 x 3x 4 y 3 y 3x 2 y t 2 7. x(0) y(0) 0 8. x 6 x 7 y 10 9. y x 2 y 2e t 4. x 3x 4 y sin t x(0) 1, y(0) 0 y 6x 5 y 10. 5. x x 5 y cos 2t y x y 11. 6. x 2 x 4 y 2 x(0) 1, y(0) 1 y x 2 y 3 12. x 2 x 3 y 5 y 2 x y 2t x 4 x y e t y 6x y et x(0) y(0) 1 x 9 x y 2e t y 8x 2 y tet x x 5 y 2 sin t y x y 3 cost x x 2 y y 2 x y e t sin t x x y 2t y x y 2t b. Find the fundamental matrix of each of the systems below and then find a solution satisfying the given initial conditions. 1. 2 1 3 x, x(0) x 1 2 2 2. 2 1 2 x, x(0) x 4 2 1 3. 3 1 1 x, x(0) x 1 1 0 4. 3 2 1 x, x(0) x 3 9 1 152 5. 7 5 2 x, x(0) x 3 4 0 6. 0 6 5 2 x 2 1 2 x, x(0) 1 4 2 4 0 7. 2 2 3 1 x 5 4 2 x, x(0) 0 5 1 5 3 c. In each of the problem below apply the method of variation of parameters to find a particular solution of the given system. 7. 2 3e 2t 1 x x 2 2 5 t 2. 4 1 e t x t x 5 2 2e 8. 3 1 t12 x 1 x 9 3 t 3 3. 2 4 cos3t x x 5 2 sin 3t 9. 3 5 cos 4t x x 5 3 sin 4t 4. 2 1 sect x x 5 2 tan t 10. 2 1 e 2t tan t x x 1 2 0 5. 0 3 2 x 3t x 3 e cos 2t 2 11. 2 4 ln t x x 1 2 t 6. Find the particular solution for each of the following above such that 1. 6 7 2t x x 1 2 3 x1 (1) x2 (1) 0 . Summary - Normal Form The normal form in the general case of a linear system of n differential equations in n unknown functions x1 , x2 ,, xn is: 153 x1 a11 (t ) x1 a12 (t ) x2 a1n (t ) xn f1 (t ) x2 a21 (t ) x1 a22 (t ) x2 a2n (t ) xn f 2 (t ) xn an1 (t ) x1 an 2 (t ) x2 ann (t ) xn f n (t ) real - Solution To A Linear System Of Differential Equations The solution to the system: dx1 f1 t , x1 , x2 , x3 , xn dt dx2 f 2 t , x1 , x2 , x3 , xn dt dxn f n t , x1 , x2 , x3 , xn dt is the parametric vector x1 (t ) x (t ) x 2 on open real interval a t b for which the xn (t ) system is true and the vector x must be continuous on the open - Matrix Form The Matrix Notation form of the normal is given as: dx1 dt a dx 2 11 dt a21 dx a 3 31 dt an1 dxn dt a12 a22 a32 a1n f1 t a2n f 2 t a3n f3 t ann f n t a13 a23 a33 an 2 an3 can simply be written as: dx Ax F(t ) ………….(I) dt 154 which is a NON – HOMOGENEOUS FORM. where F(t ) is the non – homogeneous term or the force term. If F(t ) 0 , that is f1 (t ) f 2 (t ) f 3 (t ) f n (t ) 0 for the real interval a t b then we have dx Ax ………………(II) dt Equation (II) is known as the HOMOGENEOUS FORM. - Solution of Homogeneous Equation In outline, the eigenvalue method for solving system x Ax goes as follows: (IV) Solve the characteristic equation det(A I) A I 0 for the eigenvalues 1 , 2 ,, n of matrix A . (V) We attempt to find n linearly independent eigenvector v1 , v 2 ,, v n (so that v k 0 for each k ) associated with these eigenvalues. That is A Iv k 0 for each k . This will not always be possible, but when it is, we get n linearly independent solutions: x1 (t ) v1e t , x 2 (t ) v 2 e t ,, x n (t ) v n e t (VI) The general solution is then a linear combination of these n solutions : x(t ) c1 v1e t c2 v 2 e t cn v n e t - Non – Homogenous The Method of Solving Non – Homogeneous systems 7. Solve for the homogeneous part dx Ax . dt 8. Form your fundamental matrix Φ(t ) 9. Find Φ1 (s) 155 Φ t 1 (s)F(s)ds 10. Evaluate 11. Find x p Φ(t ) Φ(s)F(s)ds . t0 t t0 Your final step x(t ) Φ(t )Φ 1 (t 0 )b Φ(t ) Φ 1 (s)F(s)ds . t t0 156 UNIT SIX SOLUTION SERIES 2.0 INTRODUCTION We have seen that a linear differential equation with constant coefficients can be reduced to the algebraic problem of finding the roots of its characteristic equation. There is no similar procedure for solving linear differential equations with variable coefficients, (at least not routinely and in finitely many steps) with the exception of special types, and the occasional equation that can be solved by inspections (perhaps followed by reduction of order). For linear ordinary differential equations with variable coefficients and of order greater than one, probably the most generally effective method of attack is that based upon the use of power series. To simplify our work and statements, the equations treated (that is the ones to be solved completely) here will by restricted to those with polynomial coefficients. The difficulties to be encountered, the methods of attack, and the results accomplished all remain essentially unchanged when the coefficients are permitted to be functions that have power series expansion valid about some point (such function are called analytic functions). 2.1 SERIES SOLUTIONS TO LINEAR DIFFERENTIAL EQUATION There are two standard methods of solving linear differential equation in series form: 1. Power Series Method 2. Frobenius Method (which is an extension of Power Series Method). Recall the theory of power series: We first remember that a power series (in powers of x x0 ) is an infinite series of the form: 157 an x x0 n a0 a1 x x0 a2 x x0 2 ak x x0 k (2.1 n 0 where a0 , a1 , a2 , are constants, called the coefficients of the series, x0 is a constant, called the centre of the series, and x is a variable. If in particular x0 0 , we obtain a power series in powers of x as: an x n a0 a1 x a2 x 2 ak x k _____ (2.2) n 0 we assume that all variables and constant are real. Recall typical Maclaurin’s series: ex 1 x x 2 x3 xk xn …………………………. (I) 1 2! 3! k! n0 n sin x x 1k x 2k 1 1n x 2n1 ….. (II) x3 x5 x7 2k 1! 3! 5! 7! n0 2n 1! cos x 1 1k x 2k 1n x 2n ………. (III) x2 x4 x6 2n! 2k ! 2! 4! 6! n0 cosh x 1 x2 x4 x6 x 2k x 2n ……………… (IV) 2k ! 2! 4! 6! 2 n ! n0 x3 x5 x7 x 2k 1 x 2n1 ……… (V) sinh x x 2k 1! 3! 5! 7! n0 2n 1! x 2 x3 1k 1 x k 1n1 x n ………….. (VI) ln1 x x 2 3 k! n! n0 1 1 x x 2 x 3 x k x n ………………………….. (VII) 1 x n 0 Power series such as those listed above are often derived as Taylor Series. The Taylor series with centre x x0 of the function f is the power series: n0 f n x0 x x0 n f x0 f x0 x x0 f x0 x x0 2 n! 2! 158 in powers of x x0 , under the hypothesis that f is infinitely differentiable at x0 (so that the coefficients of the power series are all defined). If the Taylor series of f converges to f x for all x in some open interval containing x0 , then we say that the function is analytic at x x0 or we say converges to f x in some neighbourhood of x0 . x Polynomials, sin x , cos x and e are analytic everywhere; so too are sums, difference and products of these functions. Quotients of any two of these functions are analytic all powers where the denominator is not zero. Note: 0! 1 2.2 SHIFT OF INDEX OF SUMMATION OF INFINITE SERIES 1. The index of summation of an infinite series as a dummy parameter that is: 2n x n 2k x k 2p xp n0 n! k 0 k! p0 p! 2. Indices used to maintain a particular generic term in an infinite series. Example 2.1 a n x n a 2 x 2 a3 x 3 a 4 x 4 n2 ar 2 x r 0 (ii) Consider r 2 a0 2 x 0 2 a1 2 x1 2 a22 x 2 2 an 2 x n 2 n 0 n 2n 1an x x0 n2 and maintain the generic n2 term x x0 in the infinite series: n n 2n 1a x x n2 n2 n 0 43a2 x x0 0 54a3 x x0 65a4 x x0 2 4 03 0a20 x x0 0 4 13 1a21 x x0 4 23 2a22 x x0 2 159 4 n3 na2n x x0 n n 0 (iii) x 2 n r an x nr 1 and maintain the generic term x nr . n 0 n 0 n 0 x 2 n r an x nr 1 n r x nr 1 0 r a0 x 0r 1 1 r a1 x1r 1 2 r a2 x 2r 1 n r 1an1 x nr n1 (iiii) Solve the expression n 1 n 0 nan x n1 an x n Solution: we shift index of summation of the first series, then we get: n 0 n 0 n 1an1 x n an x n . n 1an1 an x n 0 , which is true for all x . n 0 Then n 1an1 an 0 n 1,2,3, an1 an n 1 Hence for a0 0 a1 a 0 a0 1 1! a2 a1 a0 a0 2 1 2 2! a3 a2 a0 a0 3 2!3 3! 160 Thus allows the determination of: an If asked to find a0 n! n . an x n , with n0 a n so determined we have: a0 n xn n! x a0 n! a0e x . n0 n0 Exercise 2.1 With (iv) determine a n so that n1 n 0 nan x n1 2 an x n 0 2. 1. n1 n 0 nan x n1 2 an x n 0 Theorem 2.1: IDENTITY PRINCIPLE If n 0 n 0 an x x0 n bn x x0 n for every point x in some open interval I , Then an bn for all n 0 . 2.3 THE POWER SERIES METHOD Consider the general inhomogeneous linear differential equation of n th order as: bn ( x) dny d n1 y d n 2 y dy b ( x ) b ( x ) b1 ( x) b0 ( x) y f ( x) n1 n 2 n n1 n2 dx dx dx dx Then the second – order inhomogeneous equation would be: b2 ( x) d2y dy b1 ( x) b0 ( x) y f ( x) ………….. (2.3) 2 dx dx or b2 ( x) y b1 ( x) y b0 ( x) y f ( x) ………………… (2.3) The idea of the power series method for solving differential equations is simple natural. For (2.3) b2 ( x) y b1 ( x) y b0 ( x) y f ( x) 161 y b ( x) b1 ( x) f ( x) y 0 y b2 ( x) b2 ( x) b2 ( x) This can be simplified as: y p( x) y q( x) y g( x) ………..…… (2.4) where p( x) b ( x) b1 ( x) f ( x) and g( x) . , q( x) 0 b2 ( x) b2 ( x) b2 ( x) For a given homogeneous linear equation of the second – order of (2.4) That is y p( x) y q( x) y 0 …………………….. (2.5) we first represents px and qx by power series in powers of x (or x x0 , if solutions in power of x x0 is what we needed). Let y a0 a1 x a2 x 2 a3 x 3 ak x k an x n ………………………..(a) n 0 y a1 2a2 x 3a3 x 2 kak x k 1 nan x n1 ……………........(b) n1 y 2a2 3 2a3 x 4 3a4 x 2 k k 1ak x k 2 nn 1an x n2 (c) n1 Then we substitute (a), (b) and (c) into (2.5). The coefficient a0 , a1 , a2 , of an analytic solution yx an x n (i.e. (a)) of n 0 equation (2.5) can be computed as in the following steps; Step 1: Insert yx an x n into the differential equation [ or (a), (b), (c) into n 0 (2.5)]. Step 2: Rearrange the left – hand side of the equation so that the equation has the form cn x n 0 where cn is an equation in terms of n 0 Step 3: a j terms. By theorem (2.1) cn 0 for every n . Solve this equation for the a j term having the largest subscript. The resulting equation is known as the Recurrence formula for the given differential equation. 162 Step 4: Use the Recurrence formula to sequentially determine a j j 2,3, in terms of, a0 and a1 if it is second order differential equation, and a0 if it is first – order. Example 2.2 1. Solve the linear differential equation y y 0 . Solution: Let y a0 a1 x a2 x 2 a3 x 3 ak x k an x n . n1 y a1 2a2 x 3a3 x 2 kak x k 1 nan x n1 n1 Substituting these expressions into the equation, we get: n1 n 0 nan x n1 an x n 0 we shift index in the summation of first sum, i.e. replacing n 1 by n in the summation to maintain the generic term x n , we have n 1an1 x an x n 0 n n 0 n 0 n 1an1 an x n 0 n 0 by Identity principle: n 1an1 an 0 . an an , n 1 which is a recursion formula that gives as the following: a1 a0 a a a a a a , a2 1 0 0 , a3 2 0 0 ,…, 1 2 1 2 2! 3 3 2! 3! Continuing the processes, then: Therefore, a0 n x n0 n! y an x n n 0 an 163 a0 n! xn n0 n! y a0 The solution is given as: y a0 e x , where a0 is an arbitrary constant. Alternatively, y y 0 a 1 2a2 x 3a3 x 2 a0 a1 x a2 x 2 0 Then we collect like powers of x , finding a1 a0 2a0 a1 x 3a3 a2 x 2 0 Equating the coefficient of each power of x to zero, we have: a1 a0 0, 2a0 a1 0, 3a3 a2 0 Solving these equations, we may express a1 , a2 , a3 , in terms of a0 , which remains arbitrary: a1 a0 , a2 a a1 a0 a , a3 2 0 ,… 2 2! 3 3! with these coefficients the power series: y a0 a1 x a2 x 2 ak x k becomes: y a0 a0 x a0 2 a0 3 a x x 0 xk 2! 3! k! and we see that we have obtained the familiar general solution: xn a0 e x , where a0 is arbitrary. n ! n0 y a0 2. Solve y 2xy 0 . Solution: y 2xy a1 2a1 x 3a3 x 2 2 x a0 a1 x a2 x 2 a1 2a2 x 3a3 x 2 2a0 x 2a1 x 2 2a2 x 3 equating coefficients, we see that: 164 a1 0, 2a2 2a0 , 3a3 2a1 , 4a4 2a2 , 5a5 2a3 , … hence coefficients with odd subscript, a3 0 , a5 0 ,… and for the coefficients with even subscripts: a2 a0 , a4 a a2 a0 a , a6 4 0 , 2 2! 3 3! a0 remains arbitrary. With these coefficients the power series gives the following solution, which you should conform by the method of separating variables: 2 x4 x6 2 y a0 1 x a0 e x 2! 3! Alternatively, Solve y 2xy 0 Solution: n 0 n 1 Let y a n x n , then y' nan x n 1 By substitution: n 1 n 0 nan x n1 2 x an x n 0 n 1 n 0 nan x n1 2 an x n1 0 For the same generic term, we shift the left index by 2 , to maintain n 1: n 1 n 0 n 2an2 x n1 2 an x n1 0 Since the two summations do not start from the same point (i.e either n 0 or n 1). We let them start from the same point. We then find the term n 1(that is the left summation, which the start with n 0 ): n 0 n 0 1 2a1 2 x 11 n 2an 2 x n1 2 an x n1 0 165 Group like terms: 1a1 x 0 n 2an 2 2an x n1 0 n 0 Comparing coefficients, we get: a1 0 , and n 2an2 2an 0 an2 2a n n 2 Which is the recursion formula for n 0,1,2, a1 0 a5 2a 3 20 0 3 2 5 a2 2a 0 2a 0 a0 0 2 2 a6 2a a 2a4 0 0 4 2 2!6 3! a3 2a1 20 0 1 2 3 a7 2a5 20 0 5 2 5 a4 2a a 2a 2 0 0 2 2 4 2! a8 2a 6 2a a 0 0 6 2 3!8 4! then we have: if n is odd 0 a an 0 if n is even n2 ! a2 a0 , a4 or a2n a0 n! a a2 a0 a , a6 4 0 , 2 2! 3 3! Hence by substitution: 2 x4 x6 y a0 1 x 2 a0 e x 2! 3! 3. Solve y y 0 . Solution: Let y an x n n 0 n2 n 0 Then we get: nn 1an x n2 an x n 0 166 Shifting index in the first summation, and maintaining the generic term xn . n 0 n 0 n 2n 1an2 x n an x n 0 n 2 n 1 an2 an xn 0 n 0 Now, by theorem 2.1: n 2 n 1 an 2 an 0 an2 an n 2 n 1 Which is the recursion formula for n 0,1,2, a2 a0 2! a3 a1 a1 3 2 3! a4 a2 a0 a0 4 3 4 3 2! 4! a5 a3 a1 a1 5 4 5 4 3! 5! a6 a0 a a4 0 6 5. 6 5 4! 6! a7 a5 a1 a 1 7 6 7 6 5! 7! Since, y a0 a1 x a2 x 2 ak x k Thus, a a a a y a0 a1x 0 x 2 1 x3 0 x 4 1 x5 3! 5! 2! 4! x 2 x 4 x6 y a0 1 2! 4! 6! x3 x5 x 7 a1 x 3! 5! 7! n 1 n x 2 n1 1 x2n y a0 a1 n0 2n ! n0 2n 1! y a0 cos x a1 sin x 167 Exercise 2.2 In each of 1-16, find a power series solution of the given differential equation. Determine the radius of convergence of the resulting series. 1. y 2 y 0 9. 2 x 1 y y 2. y y 10. x 1 y 2 y 0 3. 2 y 3 y 0 11. 2 x 1 y 3 y 4. y 2xy 0 12. x 3 y 2 y 0 5. y x2 y 13. xy 3 y 0 6. x 2 y y 14. x 2 y 2 xy 7. y 4 y 15. 8. 2 x 1 y 2 y 0 16. 1 x y 2xy xy x 1 y 2 2 Exercise 2.3 Fine two linearly independent power series solutions of to given differential equation. Determine the radius of convergence of each series, and identify the general solution in terms of familiar elementary functions. 1. y y 7. y 4 y 2. y 9 y 0 8. y y x 3. y 4 y 0 4. y 4 xy 4 x 2 2 y 0 5. y 3 y 2 y 0 6. 1 x y 2xy 2 y 0 . 2 168 2.4 SOLUTIONS ABOUT NEIGHBOURHOOD POINTS The power series method is used to solve linear differential equations such as y p x y q x y g ( x) ………………………. (2.4) where p x and q x are given function as if x x0 solutions obtained are power series solution in the neighbourhood of x x0 . To obtain this solution p x and q x need to satisfy the following; (1) x x0 is an ordinary point if both of the following limits exist and lim p x , lim q x are finite: x x0 i.e.: (2) xx0 lim p x , lim q x . x x0 x x0 x x0 is singular point if either lim p x , lim q x x x0 x x0 i.e. they are not finite (infinite) a) x x0 is a regular point if it is a singular point and if both of the following limits exist and are finite: lim x x0 p x , lim x x0 q x 2 x x0 xx0 i.e. lim x x0 p x , lim x x0 q x 2 x x0 xx0 b) x x0 is a irregular singular point if either (or both) of limits do not exist. lim x x0 p x , lim x x0 q x 2 x x0 xx0 i.e. they are not finite (infinite). NOTE: 169 The solutions we have obtained from the examples of Power series is about the point x 0 (i.e x0 0 ) p 0 and q 0 are ordinary point. Example 2.3 1. ☞ The differential equation 1 x y 6xy 4 y 0 2 has x 1 and x 1 as its only singular points in the finite plane. ☞ The equation y 2xy y 0 has no singular points in the finite plane. ☞ The equation xy y xy 0 has the origin x 0 as the only singular point in the finite plane. 2. Determine whether x 0 is an ordinary point of the differential equation y xy 2 y 0 . Here p x x and q x 2 are both polynomials; hence they are analytic everywhere. Therefore, every value of x , in particular x 0 , is an ordinary point. 3. Determine whether x 0 is an ordinary point of the differential equation y y 0 . Here p x 0 and q x 1 are both constants; hence they are analytic everywhere. Therefore, every value of x , in particular x 0 , is an ordinary point. 4. Determine whether x 0 is an ordinary point of the differential equation x2 y 2 y xy 0 . 170 Here p x 2 1 and q x . Neither of these functions is analytic at x 0 , 2 x x so x 0 is not an ordinary point, rather, a singular point. 5. Determine whether x 0 or x 1 is an ordinary point of the differential equation (called Legendre’s equation) 1 x y 2xy n(n 1) y 0 2 for any positive integer n . Hence p x n n 1 2 x and q x ; 2 1 x2 1 x If x 0 , both are analytic there and x 0 is an ordinary point. If x 1 , neither function is analytic there. Consequently, x 1 is a singular point. 6. Classify the singular points, in the finite plane, of the equation x 1 x x 2 y x2 y ( x3 2x 1) y 0 2 Hence p x x x 12 x 2 and q x x3 2 x 1 x x 1 x 2 2 ; The singular points in the finite plane are x 0,1, 2 . Consider x 0 : The factor x is absent from the denominator of p x and it appears to the first power in the denominator of q x . Hence x 0 is a regular singular point. Consider x 1 : The factor x 1 appears to the second power in the denominator of p x . That is a higher power than is permitted in the definition of a regular singular 171 point. Hence it does not matter how x 1 appears in q x ; the point x 1 is an irregular singular point. Consider x 2 The factor x 2 appears to the first power in the denominator of p x , just as high as is permitted, and to the first power also in the denominator of q x , so x 2 is a regular singular point. In summary, the equation has in finite plane the following singular points: 7. ~ Regular singular points at x 0, x 2 . ~ Irregular singular points at x 1 . Classify the singular points in the finite plane for the equation x4 x2 1 x 1 y 4x3 x 1 y ( x 1) y 0 Hence p x q x 2 4 x x 1 x 1 2 4 and x x i x i x 1 x 1 x x i x i x 1 4 2 Therefore the desired classification is; Regular Singular Point (R.S.P) at x i, i,1 and Irregular Singular Point (I.S.P) at x 0 8. In the equation sin 2x y x x 1 y x2 y 0 there are infinite number of singular points, namely, x , , 3 , 2 2 Note that x 0 is not a singular point, 172 since lim x0 x x 1 1 x2 and lim 0 x0 sin 2 x sin 2 x 2 Exercise 2.4 For each equation, list all of the singular points in the finite plane. 1. x 2. x 3 x y 3 x y 4 xy 0 12. x2 x2 9 y 3xy y 0 3. 4 y 3xy 2 y 0 x2 4. x x 1 y 3xy x 1 y 0 14. 5. x2 y xy 1 x2 y 0 15. 2 x 1 x 3 y y 2 x 1 y 0 6. x4 y 0 16. x3 x 2 4 17. x x2 1 y y 0 18. x 19. 4 x 1 y 3xy y 0 2 4 y 6 xy 3 y 0 11. 13. 2 8. 1 x y 2xy 6 y 0 x 4x 3 y x y 4 y 0 9. x2 1 x y 1 2x y 0 10. 6 xy 1 x2 y 2 y 0 7. 2 2 2 3 4 y y 0 1 4x y 4xy y 0 2 4xy y 0 2 2 y 2 x 2 4 y xy 0 6 x 8 y 3 y 0 Exercise 2.5 For each equation, locate and classify all its singular points in the finite plane. 1. x3 x 1 y x 1 y 4xy 0 10. 2. x2 x2 4 y 2 x3 y 3 y 0 3. y xy 0 12. x 1 x 2 y 5 x 2 y x2 y 0 4. x2 y y 0 13. x 12 x 42 y x 4 y 7 y 0 5. x4 y y 0 14. 1 4x y 6x 1 4x y 9 y 0 11. 2 173 x2 x 2 y x 2 y 4 y 0 x x 3 y y y 0 2 6. x 1 x 4 7. y x 4 y y 0 15. x3 y 4 y 0 x2 x 2 y 3 x 2 y y 0 16. 1 4x y 6xy 9 y 0 8. x2 x 4 y 3xy x 4 y 0 17. x4 y 2x3 y 4 y 0 9. 1 4x 2x 14 y 2x 1 y 8 y 0 3 2 2 2 2 2 y 6 xy 9 y 0 18. 2 Exercise 2.6 Determine and classify the singular points (if any) of each of the following equation. 3. x x x 4. sin 2xy x 3 y e x y 0 5. 3x2 x 2 y 2x x 1 y x2 2 y xe x 1. 2. 6. 7. 8. 3 7 x 8 y x 3 y xy 0 4 16 y x 2 y x 2 y 0 4 4 x2 4 y xy 0 sin 3x y x x 2 y 5xe y 2e 9. 2 2 x 2x x x 1 y sin 2x y 4x y sin 2x 1 cos x y 3xy 4 y 0 10. x sinh x y 3sin x y 14e y 0 e 1 y 14 y 12e cos x y 0 2 2 1 2 x 2 x x Examples of Solution about a neighbourhood point x x0 1. Find the general solution near x 0 (i.e. x0 0 ) of y xy 2 y 0 . Solution: Here p x x and q x 2 . x 0 is an ordinary point since p( x) and q x are polynomials that means it is analytic everywhere. Let y an x n n 0 y nan x n1 n1 Therefore, 174 y n n 1 an xn2 n2 2a2 6a3 x 12a4 x2 n n 1 an x n2 n 1 nan1x n1 n 2 n 1 an2 xn x a1 2a2 x 3a3 x 2 4a4 x3 nan x n1 n n1 n 1 an1x n 2 an2 x 2 a0 a1x a2 x2 a3 x3 an xn an1xn1 0 Combining terms that contain like powers of x , we have: 2a2 2a0 x 6a3 a1 x2 12a4 x3 20a5 a3 xn n 2 n 1 an2 nan 2an 0 0x 0x2 0xk The last equation holds if and only if each coefficient in the left – hand sides is zero. Thus: 2a2 a0 0,6a3 a1 0,12a4 0,20a5 a3 0 In general: n 2 n 1 an2 n 2 an 0 , or an2 n 2 a n 2 n 1 n which is the recurrence formula for this problem. Calculate a 2 a 0 a4 0 1 a3 a1 6 1 1 1 1 a5 a3 a1 a1 20 20 6 120 1 1 1 a7 a1 a1 14 120 1680 2 1 a4 0 30 15 4 1 a8 a6 0 0 a9 ..... 56 14 a6 Note that since a4 0 , it follows from the recurrence formula that all the even coefficients beyond a4 are also zero. Substituting we have 175 1 1 1 y a0 a1x a0 x2 a1x3 0 x 4 a1x5 0 x6 a1x7 6 120 1680 1 1 5 1 7 y a0 1 x 2 a1 x x3 x x 6 120 1680 2. Solve the equation y x 1 y 4 x 1 y 0 about the ordinary point 2 x 1. Solution: To solve an equation “about the point x x0 ” means to obtain solution valid in a region surrounding the point, solutions expressed in powers of x x0 . We first translate the axes, putting x 1 v . The equation then becomes d 2 y 2 dy v 4vy 0 dv dv2 always in a pure translation, x x0 v , we have dy dy , and so on. dx dv As usual we put : y anv n n 0 and we obtain: n2 n1 n 0 n n 1anvn2 nanvn1 4 anvn1 0 collecting like terms yields: n2 n 0 n n 1anvn2 n 4 anvn1 0 which, with a shift of index from n to n 3 in the second series, gives: n 0 n 3 n n 1anvn2 n 7 an3vn2 0 Therefore a0 and a1 are arbitrary and for the remainder we have n 2: 2a2 0, n 3: n n 1 n 7 an3 0 176 an n7 an3 n n 1 4 a0 3 2 1 a6 a3 65 2 a9 a6 9 8 3 a1 43 0 a7 a4 0 76 3 a10 a7 0 10 9 a3 a4 2 a2 0 5 4 a7 a5 0 a5 a11 0 a3n2 0, n 1 a3n1 0, n 2 3n 7 a3n a3n3 3n 3n 1 with the usual multiplication scheme, the first column yields: n 1: 1n 4 1 2 3n 7 a0 a3n 3 6 9 3n 2 5 8 3n 1 1n 4 1 2 3n 7 v3n a1 v 14 v4 y a0 1 n1 3 6 9 3n 2 5 8 3n 1 since v x 1 , the solution appears as: 1n 4 1 2 3n 7 x 13n a1 x 1 14 x 14 y a0 1 n1 3 6 9 3n 2 5 8 3n 1 but 3 6 9 3n 3n 1 2 3 n 3n n! Furthermore, all but the first two factors inside the square bracket in the numerator also appear in the denominator, and from n 2 , they cancelled out, it can be shown that : 4 1 x 1 n y a0 3n 4 a1 x 1 14 x 1 n0 3 3n 1 3n 4 n ! n 177 Exercise 2.7 1. Find the general solution in powers of x of x2 4 y 3xy y 0 . Then find the particular solution with y 0 4, y 0 1 2. Determine the general series solution of the equation x 1 y xy 4 y 0 about the point x 0 . 2 3. Find the general solution near x 2 of y x 2 y 2 y 0 . 4. y 2 x 3 y 3 y 0 . Solve about x 3 . 5. y x 2 y 0 . Solve about x 2 6. Find the general solution of the Legendre’s equation 1 x y 2xy n n 1 y 0 2 2.5 THE FROBENIUS METHOD The standard form of linear homogeneous differential equation of second order of equation (2.5) is: y p( x) y q( x) y 0 ………….(2.5) We shall confine our discussion to a solution of (2.5) in the neighbourhood of x 0 . The Frobenius method applicable when x 0 is ordinary or regular point of (2.5). Any differential equation of the form: y p x q x y 2 y 0 ………(2.6) x x where the functions p x and q x are analytic at x 0 , we use a more general method to assume a trial solution of the form: y xr a0 a1x a2 x2 a3 x3 x a x r n 0 n where a0 is the first coefficient that is not zero (i.e. a0 0 ). 178 n …(2.7) Indicial Equation Assume a series solution of the form: y xr a0 a1x a2 x2 a3 x3 x a x r n 0 n n ……..(2.7) a0 0 of the differential equation x2 y xp( x) y q( x) y 0 …………………………(2.6a) we first expand p x and q x in power series: q x q0 q1x q2 x2 q3 x3 p( x) p0 p1x p2 x 2 p3 x3 Termwise differentiation of trial solution in (2.7) y x xr 1 ra0 r 1 a1x n r an x nr 1 n 0 y x xr 2 r r 1 a0 r r 1 a1x n r n r 1 an x nr 2 n 0 by inserting all these series into (2.6a) we readily obtain: xr r r 1 a0 q0 q1x xr ra0 r 1 a1 ..(2.8) q0 q1x xr a0 a1x 0 we now equate the sum of the coefficients of each power xr , xr 1, xr 2 , to zero. This yields a system of equations involving these unknown coefficients an . The equation corresponding to the power xr is r r 1 p0r q0 a0 0 …….(2.9) Since by assumption a0 0 , the expression in the brackets must by zero. This gives r r 1 p0 r q0 0 ……………..(2.10) This important quadratic equation is called indicial equation of the differential equation (2.7). This is the coefficient of the lowest power of x gives the indicial equation from which values of r are obtained, r r1 and r r2 , the roots of the equation. 179 Three Cases Arises Case 1: Distinct roots not differing by an Integer. A basis is y x x c y1 x xr1 a0 a1x a2 x2 and 2 r2 0 c1x c2 x2 …………(I) …………(II) with coefficients obtained successively from (2.8) with r r1 and r r2 , respectively. Case 2: Double root r r1 r2 . A basis is y1 x xr a0 a1x a2 x 2 ……….(III) r 12 1 p0 and …..(IV) y2 x y1 x ln x xr c1x c2 x 2 x0 Case 3: Roots differing by an integer. A basis is ………..(V) c c x c x …….(VI) y1 x xr1 a0 a1x a2 x2 and y2 x ky1 x ln x xr2 2 0 1 2 where the roots are so denoted that r1 r2 0 and k may turn out to be zero. 180 Example 2.5 Solve the equation 4 x d2y dy 2 y 0. 2 dx dx Solution: y 1 1 y y0 2x 4x clearly p x 1 1 , q x . 2x 4x Therefore the problem makes the point x 0 a regular singular point. Let y an x nr n 0 y n r an x nr 1 n 0 y n r n r 1 an x nr 2 n 0 substituting the above into the differential equation we get: n 0 n 0 n 0 4 n r n r 1 an x nr 1 2 n r an x nr 1 an x nr 0 n 0 n 0 2 n r 2n 2r 1 an x nr 1 an x nr 0 n1 n 0 2r 2r 1 a0 xr 1 2 n r 2n 2r 1 an x nr 1 an x nr 0 n 0 n 0 2r 2r 1 a0 xr 1 2 n r 1 2n 2r 1 an x nr an x nr 0 2r 2r 1 a0 xr 1 2 n r 1 2n 2r 1 an1 an x nr 0 n 0 This last expression is true for all x , therefore the coefficients of all powers of x must vanish. The coefficient of the lowest power of x , 2r 2r 1 a0 is called the indicial term and it determines the values of the index r . 181 In this case i.e. Case 1, assuming a0 0 then r 0, 12 . The other coefficients of x n r for n 0,1,2,3 lead to the recurrence relationship: 2 n r 1 2n 2r 1 an1 an 0 an1 an 2 n r 1 2n 2r 1 2r 2r 1 a0 0 …….. indicial equation If r 0 , a0 0 , then the recurrence relation of coefficient becomes: an1 an 2 n 1 2n 1 if n 0 , a1 a0 a 0 2 11 2! if n 1 , a2 a a1 a 1 0 2 23 4 3 4! if n 2 , a3 a a2 a 2 0 2 35 6 5 6! if n 3 , a4 a3 a a 3 0 2 4 7 8 7 8! therefore n 1 a0 an 2n ! Recall that (I), y1 x r1 an x n . n 0 x0 a0 a1x a2 x 2 ak x k x x2 x3 a0 1 2! 4! 6! a0 1 k 1 xk 2k ! x x x 2 2! 4 4! 6 6! 1k y a0 cos x , a0 is arbitrary constant 182 2! x 2k The other value obtained for r 12 . Recall the recurrence relation: cn1 if n 0 , c1 if n 1 , c2 if n 2 , c3 Therefore, cn c is arbitrary, 2 n 2n 2 0 3 2 c0 c 0 3 2 3! 2 c1 5 2 4 c2 2 7 2 6 c1 c0 5 4 5! c2 c0 7 6 7! n 1 c0 cn 2n 1! Recall that (II), y2 xr2 cn x n n 0 1 x 2 c0 c1x c2 x2 x x 2 x3 x c0 1 3! 5! 7! 1 2 c0 x ck xk k 1 xk 2k 1! x x x 3 3! y2 c0 sin x , Hence for y y1 y2 The general solution is given as: y a0 cos x c0 sin x or. 183 5 5! 7 7! c0 is arbitrary 1k k! x k Example 2.6 Find the exponents in the possible Frobenius series solutions of the equation 2x2 1 x y 3x 1 x y 1 x2 y 0 3 Solution: We divide each term by 2 x2 1 x to recast the differential equation in the form y 32 1 2x x2 x and thus see that p0 3 2 y 12 1 x y 0 x2 and q0 12 Hence the indicial equation is: r r 1 32 r 12 r 2 12 r 12 r 1 r 12 0 with roots r1 1 2 and r2 1. The two possible Frobenius series solutions are then of the forms n 0 n 0 y1 x 2 an x n and y2 x1 cn x n . 1 Example 2.7 Find a Frobenius series solution of Bessel’s equation of order zero: x2 y xy x2 y 0 Solution: y 1 x2 y 2 y 0 x x Hence x 0 is a regular singular point with p x 1 and q x x2 , so our series will converge for all x 0 . Because p0 1 and q0 0 , the indicial equation is r r 1 r r 2 0 Thus we obtain only the single exponent r 0 and so there is only one Frobenius series solution 184 y x0 an x n n 0 substitute in the differential equation; the result is: n 0 n 0 n 0 n n 1 an xn nan xn an xn2 0 we combine the first two sums and shift the index of summation in the third by 2 to obtain: n 0 n2 n2an xn an2 xn 0 Therefore, the recursion formula is: an an2 n2 for n 2 a2 a0 , 22 a4 a a a2 a 2 0 2 and a6 42 2 02 2 2 4 24 6 246 n n 1 a0 1 a0 a2n 2 2 22 42 2n 22n n! The choice a0 1 gives us one of the most important special functions in mathematics, the Bessel function of order zero of the first kind, denoted by J0 ( x) . Thus n 1 x2n x2 x4 x6 J 0 ( x) 1 2 2n 4 64 2304 n0 2 n! In this example we have not been able to find a second linearly independent solution of Bessel’s equation of order zero. 185 Exercise 2.9 In each exercise, determine whether x 0 is an ordinary point, a regular singular point, or an irregular singular point. If it is a regular point, find the exponent of the differential equation at x 0 . 1. xy x x3 y sin x y 0 1a. xy x2 y e x 1 y 0 2. x2 y cos x y xy 0 2a. x2 y 6sin x y 6 y 0 3. 3x3 y 2 x2 y 1 x2 y 0 4. Apply the method of Frobenius of Bessel’s equation of order 1 2 x2 y xy x2 14 y 0 , to derive its general solution for x 0 , y a0 5. cos x sin x a1 x x Show that Bessel’s equation of order 1, x2 y xy x2 1 y 0 has exponents r1 1 and r2 1 at x 0 , and that the Frobenius series 2n x 1 x solution corresponding to r1 1 is J1 x 2 n0 n! n 1!22n n 6. Obtain the solution x y y 0 by Frobenius method. 7. Find the series solutions in powers of Airy’s equation x y xy 8. Find the general series solution of the following a) 2 xy 3 2 x y 4 y 0 b) c) x2 y xy x2 1 y 0 186 2 x x2 y 1 9 x y 3 y 0 UNIT SEVEN FOURIER SERIES Definition: Periodic Function If the value of each ordinate f t repeats itself at equal intervals in abscissa, t , the f t is said to be periodic. Thus if f t f t T f t 2T ... , then T is called the least period (or period) of the function f t . Example 3.1 1: The function sin x has periods 2 ,4 ,6 since sinx 2 , sinx 4 , sinx 6 all equal sin x . So the least period is 2 . 2: The period of sin nx or cosnx , where n a positive integer, is 2 / n . 3: The period of 4: A constant has any positive number as period. tan x is . Other examples of periodic functions are shown in graphs below: period period Piecewise Smooth Functions Definition: A function f x is piecewise continuous on the open interval a x b if: 1. f x is continuous everywhere in a x b with the possible exception of at most a finite number of points x1, x2 ,..., xn and 2. at these points of discontinuity, the right- and left – hand limits of f x , respectively lim f x and lim f x , exist ( j 1,2,..., n ) x x j x x j x x j x x j 187 (Note that a continuous function is piecewise continuous). Definition: A function f x is piecewise continuous on the closed interval a x b if: 1. it is piecewise continuous on the open interval a x b , 2. the right – hand limit of 3. the left – hand limit of f x exists at x a , and f x exists at x b . Definition: A function f x is piecewise smooth on a, b if both f x and f x are piecewise continuous on a, b . Examples 3.2 1. Determine whether x2 1 x 0 f x is piecewise continuous on -1,1 1/ x x 0 Solution: 1,1 except at x 0 . Therefore, if the right- and left- hand limits exist at x 0 , f x will be piecewise continuous on 1,1 . We have: The given function is continuous everywhere on lim f x lim x2 1 1 x0 x0 x0 x0 1 lim f x lim x0 x0 x x 0 x 0 Since the left – hand limit does not exist, on 2. f x is not piecewise continuous 1,1 . x 1 sin x 0 0 x 1 Is f x x piecewise continuous on -2,5 ? e 1 x 0 x3 x 1 188 Solution: The given function is continuous on 2,5 except at the two points x1 0 and x2 1.(Note that f x continuous at x 1 ). At the two points of discontinuity, we find that: lim f x lim0 0 lim f x lim e x 1 lim f x lim e x e1 lim f x lim x3 1 x0 x 0 and x1 x 1 x1 since all require limits exist, 3. x0 x 0 x0 x1 x 1 x0 x1 f x is piecewise continuous on 2,5 x2 1 x<0 Is f x 1 0 x 1 piecewise smooth on -2,2? 2 x 1 x 1 Solution: The function is continuous everywhere on Since the require limits exist at Differentiating 2,2 except at x1 1. x1, f x is piecewise continuous. f x , we obtain x<0 2 x f x 0 0 x 1 2 x 1 The derivative does not exist at x1 1 but is continuous at all other points in 2,2 . At x1 the required limits exist; hence f x is piecewise continuous. It follows that 4. f x is piecewise smooth on 2,2 . 1 x<0 Is f x x 0 x 1 piecewise smooth on -1,3? x3 x 1 189 Solution: The function is continuous everywhere on require limits exist at 1,3 except at x1 0 . Since the x1, f x is piecewise continuous. Differentiating f x , we obtain: 0 x<0 1 f x 0 x 1 2 x x 1 3x 2 which is continuous everywhere on 1,3 except at the two points x1 0 and x2 1where the derivative does not exist. At x1 0 , lim f x lim xx1 x x1 x0 x0 1 2 x Hence , one of the required limits does not exist. If follows that piecewise continuous, and therefore that 1,3 . 190 f x is not f x is not piecewise smooth on Definition: Fourier Series f x (which is integrable in , ) can be A non – sinusoidal period function expressed in a converging series of the form: n1 n1 f x a0 an cos nx bn sin nx ………..(3.1) called the Fourier Series, where a0 , a1, a2 ,..., b1, b2 , b3 ,... are constants that can easily be determined, and f xdx a0 1 2 an f x cos nxdx, bn 1 1 f xsin nxdx n 1,2,... ……… (3.2) n 1,2,... where a0 , an , and bn are called Fourier coefficients. The function cos x and sin x are called fundamentals for n 1 an integral cosnx and sin nx are called harmonics, which are multiples of the fundamental frequencies. Orthogonality Conditions For The Sine And Cosine Functions The following integrals are very useful in the Fourier expansions: For m, n integers m n : 2 2 1. sin nxdx 0 2. 0 2 3. sin 2 nxdx 0 2 4. 0 sin nx sin mxdx 0 6. nxdx cos nx sin mxdx 0 0 2 cos nx cos mxdx 0 2 2 0 7. cos 0 2 5. cos nxdx 0 2 8. 0 sin nx cos mxdx 0 0 191 The iterated integration by parts is also very useful: uvdx uv uv 1 2 uv3 uv4 ... du d 2u d 3u , u 2 , u 3 ,... where u dx dx dx v1 vdx, v2 v1dx, v3 v2dx, v4 v3dx, … Also for N which is an integer n N , sin n 0 , cos n 1 n Determination of Fourier Coefficients n1 n1 f x a0 an cos nx bn sin nx with period 2 . Let a) Then to find a0 : Integrate both sides of (3.1) w.r.t. x for 0 x 2 2 0 2 2 2 0 0 0 f x dx a0dx 2 0 an cos nxdx n1 b sin nxdx n1 n 2 f x dx a0 dx a0 2 , other term vanishes by orthogonal 0 properties. a0 b) 1 2 2 f x dx ……….(I) 0 To find an : Multiply each side of equation (3.1) by cosnx and integrate throughout with respect to x , 0 x 2 2 0 2 2 0 0 f x cos nxdx a0 cos nxdx 2 0 an a cos 2 n1 n 2 nxdx 2 0 f x cos nxdx an cos nx cos nxdx an 0 1 2 0 f x cos nxdx ……….(II) 192 b cos nx sin nxdx n1 n c) To find bn : Multiply each side of equation (3.1) by sin nx and integrate throughout with respect to x , 0 x 2 2 0 2 2 2 0 0 0 f x cos nxdx a0 sin nxdx 2 an cos nx sin xdx n1 b sin n1 n 2 nxdx 2 f x sin nxdx bn sin nx sin nxdx bn 0 0 bn 1 2 0 f x sin nxdx ………….(III) Example and Solution If the following function are defined over the interval x and f x 2 , state whether or not each function can be represented by a Fourier series. 1. f x x3 - Yes 2. f x 4x 5 - Yes 3. f x 2 - No, infinite discontinuity at x 0 x 4. f x 1 - Yes x4 5. f x tan x - No: infinite discontinuity at x / 2 . f x y , where x2 y 2 9 - No; two valued Dirichlet Conditions Suppose that; 1. f x is defined except possibly at a finite number of points in L, L . 2. f x is periodic outside L, L with period 2L . 3. f x and f x are piecewise continuous in L, L . Then the series with Fourier coefficients converges to a) f x if x is a point of continuity b) if x is point of discontinuity. 193 Here f x 0 and f x 0 are the right – and left – hand limits of f x at x represents lim f x and lim f x respectively. 0 0 The conditions 1, 2 and 3 imposed on f x are sufficient but not necessary, and are generally satisfied in practice. Example 3.1: f x x and 0 x 2 Find the Fourier series representing the periodic function and sketch its graph from x 4 to x 4 . Solution: Clearly f x x is periodic then, n1 n1 f x a0 an cos nx bn sin nx Recall: a0 1 2 2 0 f x dx , an 1 2 0 f x cos nxdx , bn 1 2 0 f x sin nxdx Thus; 1 a0 2 2 0 an 1 2 0 bn 1 2 0 2 1 x2 xdx , 2 2 0 2 1 x 1 2 x cos nxdx sin nx sin nxdx 0 n n 0 0 2 1 x 1 2 2 x cos nxdx cos nx cos nxdx n n 0 n 0 Put the values into: n1 n1 f x a0 an cos nx bn sin nx 1 1 1 f x 2 sin x sin 2x sin3x ... sin kx ... 2 3 k 194 Sketch Function Defines in two or more sub – interval Example 1. Find the Fourier coefficients corresponding to the function 0 5 x 0 f x 3 0 x 5 period 10 2. Write the corresponding Fourier series 3. How should f x be defined at x 5 , Fourier series will converge to f x for 5 x 5 ? Solution: 1. Period 2 L 10 and L 5 . Choose the interval c to c 2 L as 5 to 5 , so that c 5 . 5 1 L 1 0 f x dx 0dx 3dx 0 2L L 2(5) 5 a0 a0 0 Then an 1 5 3 3x 25 0 2 1 c2 L n x 1 5 n x f x cos f x cos dx L c L 5 5 5 5 1 0 n x n x 3 5 n x 5 0 cos dx 0 3 cos dx 0 cos 5 5 5 5 5 3 5 n x sin 0 if n 0 5 n 5 0 5 bn 1 c2 L n x 1 5 n x f x sin f x sin dx L c L 5 5 5 195 5 1 0 n x n x 3 5 n x 5 0 sin dx 0 3 sin sin dx 5 5 5 5 0 5 31 cos n 3 5 n x cos 5 n 5 0 n 5 2. The corresponding Fourier series is: n n f x a0 an cos x bn sin L L n1 x 3 31 cos n nx sin 2 n1 n 5 3 6 x 1 3 x 1 5 x sin sin sin ... 2 5 3 5 5 5 3. f x satisfying the Dirichlet conditions, we can say that the series Since converges to f x at all points of continuity and to t at points of discontinuity. At x 5,0 and 5 , which are points of discontinuity, the series converges to 3 0 / 2 3 / 2 . If we redefine f x as follows: 3 / 2 0 f x 3 / 2 0 3 / 2 x 5 5 x 0 x0 0 x5 x5 Then the series will converges to f x for 5 x 5 . Example 3.5 Find the coefficients of the period function k f x k if x 0 if 0 x and Solution: a0 1 2 f x dx 1 0 k dx 0 kdx 0 2 196 f x 2 f x an f x cos nxdx k cos nx k cos nx 0 1 1 0 0 1 sin nx sin nx k k 0 n n 0 because sin nx 0 at bn ,0, and for all n 1,2,... f x sin nxdx k sin nx k sin nx 0 1 1 0 0 1 cos nx cos nx k k n n 0 since cos cos ;cos0 1, this yields. bn Now, k 2k cos0 cos n cos n cos0 1 cos n n n cos 1,cos2 1,cos3 1, etc; 1 for odd n cos n and thus 1 for even n In general 2 for odd n 1 cos n 0 for even n Hence the Fourier Coefficients bn of our function are b1 4k , b2 0, b3 4k , b4 0, b5 Since the an are zero, the Fourier series of 4k 1 1 sin x sin3x sin5x 3 5 197 4k 5 f x is: ODD AND EVEN FUNCTION Even Functions Definition: A function f x is said to be even (or symmetric ) if f x f x The graph of a symmetric function is symmetrical about the y axis The area under such a curve from to is Thus is also twice the area from 0 to f x dx 2 f x dx . 0 x4 ,2x6 4x2 5,cos x, e x e x are even function Odd functions Definition: A function is called odd (or skew symmetric) if f x f x . Some graphs of odd functions: The area under the curve from to Thus is zero; i.e. f x dx 0 . x3, x5 3x3 2x,sin x,tan3x are odd functions. Products of Odd and Even Functions even even even odd odd even odd even odd 198 that Expansion of the Even Function, 0 f x a0 1 2 1 f x dx f x dx an 1 f x cos nxdx bn 1 f x sin nxdx 0 2 0 f x cos nxdx Therefore the series of the even function contains only cosine terms Expansion of the Odd Function, f x a0 1 2 f x dx 0 an 1 f x cos nxdx 0 bn 1 f x sin nxdx 2 0 f x sin nxdx Therefore the series of the odd function contains only sine terms. Remark: Functions which are neither even nor odd contains both sine and cosine terms. Example 3.5 Consider the function 6 x 0 f x 6 0 x f x 2 f x Before we do any evaluation, we can see that this is an odd function and therefore sine terms only; i.e. a0 0 an 0 bn 1 f x sin nxdx f x sin nx is a product of odd functions and is therefore even. 199 12 cos nx bn f x sin nxdx 6sin nxdx 0 0 n 0 0 n even 12 1 cos n 24 n n n odd 2 2 The Fourier series is then: f x 24 1 1 sin x sin3x sin5x 3 5 Example 3.6 Obtain a Fourier expansion for f x x for x . 3 Solution: f x x3 is an odd function, therefore a0 an 0 n for bn 2 0 6 n x3 sin nxdx 2 1 3 n n 2 6 2 6 2 6 x3 2 3 sin x 3 sin 2 x 3 sin x ... 2 3 2 3 1 1 200 Half – Range Series, Periodic for 0, 0, . To get the series of cosine f x is an even function in the interval , . Let the given function be defined in the interval only, the function Then, an 2 0 f x cos nxdx and bn 0 n f x as series of sine only, extend the interval to , and treat To expand f x as odd function. Then, bn 2 0 f x sin nxdx , an 0 n Example 3.7 Find the Fourier sine series for the function f x e ax for 0 x Solution: Let eax bn sin nx n1 bn 2 0 eax sin nxdx 2n 1 1n ea a 2 n2 Thus, b1 eax 2 1 ea a 2 ,b 12 2 2 2 1 ea a 2 22 1 ea 2 1 ea sin x 2 2 2 2 2 sin 2 x ... a 1 a 2 201 Example 3.8 Expand f x x,0 x 2 in a half range a. sine series b. cosine series Solution: a) an 0 bn 2 L n x 2 2 n x f x sin dx x sin dx L 0 L 2 0 2 2 2 n x 4 n x 4 x cos cos 1 2 2 sin n 2 2 n n 0 Then 4 n x cos n sin 2 n 1 n f x 4 x 1 2 x 1 3 x sin sin sin ... 2 2 2 3 2 b) Thus bn 0 , an 2 L n x 2 2 n x f x cos dx x cos dx L 0 2 2 0 2 2 n x n x 2 4 x sin 1 cos 2 2 2 2 0 n n If 4 n 2 2 cos n 1 2 n 0, a0 xdx 2 0 Then 1 if n 0 . f x 1 n 1 n 4 2 cos n 1 cos 2 8 x 1 3 x 1 5 x cos cos cos ... 2 2 2 2 3 2 2 5 202 n x 2 Change of Integral and Functions of Arbitrary Period For some problems the period of the function is not 2 but t or 2L . To conform to the general theory, this period must be converted to its equivalent of 2 , and such will equally affect the argument of the function proportionally. Let f x be defined in the interval L, L , to convert this to 2 ; 2L is the interval for variable x 1 is the interval for the variable x 2L and therefore 2 is the interval for the variable 2 x x 2L L let z x L , then x zL zL f x of period 2L is transformed to the function f of Thus the function period 2 .Thus, zL f a0 an cos nz bn sin nz n1 n1 a0 where but an 1 x 0 1 2 0 zL 1 2L zL x f dz f x d o 2 0 L and so a0 1 2L f x dx 2L 0 1 2L n x x zL f cos nzdz f x cos d L 0 L L 2 an bn 1 2 2 0 1 2L n x f x cos dx L 0 L 1 2L n x zL f sin nzdz f x sin dx L 0 L n x n x f x a0 an cos bn sin L L n1 n1 203 In general the Fourier series with period 2L is stated as: n n f x a0 an cos x bn sin x …………….(3.3) L L n 1 with the Fourier coefficients of f x given by the Euler formulas: a0 1 L f x dx 2L L an 1 L f x cos nxdx , L L n 1,2,... bn 1 L f x sin nxdx L L n 1,2,... Corollary: For half range series in the interval f x a0 an cos Cosine series: n1 a0 where f x bn sin n1 where bn n x L 1 L 2 L n x , f x dx a f x cos dx n L 0 L 0 L Sine series: 0, L ; n x L 2 L n x f x sin dx L 0 L Example 3.9 1. Given f x Expand 0 1 0 xc c x 2c f x in the Fourier series of period 2c Solution: f x a0 an cos n1 n x n x bn sin L L n1 Recall that: a0 1 2c 1 c 1 2c 1 f x dx 0 dx 1dx 2c 0 2c 0 2c c 2 204 an 2c 1 2L n x 1 c n x n x f x cos dx 0 cos dx 1 cos dx 0 c L 0 L c 0 c c bn 2 L n x f x sin dx L 0 L bn 2c 1 2c n x 1 c n x n x f x sin dx 0 sin dx 1 sin dx c c 0 L c 0 c c 2 bn n 0 n is odd n is even 1 2 x 1 3 x f x sin sin ... 2 c 3 c 2. 0 2 x 1 1 x 1 Find the Fourier series of the function f x k 0 1 x 2 Solution: We determine the Fourier series, but the period 2c 4 c2 a0 1 c 1 2 f x dx f x dx 2c c 2 2 2 1 2 1 1 k a0 0 dx kdx 0 dx 2 1 1 4 2 an 1 2L n x 1 c n x f x cos dx f x cos dx L 0 L c c c an bn 1 2 n x 1 1 n x 2k n f x cos dx k cos dx sin 2 2 2 2 1 2 n 2 1 2L n x 1 c n x f x sin dx f x sin dx L 0 L c c c an 1 2 n x 1 1 n x f x sin dx k sin dx 0 2 2 2 2 1 2 f x k 2k 1 x 1 3 3 sin cos sin cos x ... 2 1 2 2 3 2 2 then 205 f x k 2k x 1 3 cos cos x ... 2 2 3 2 Fourier Series from the Sketches of the Wave Form Example 1: Derive the Fourier series for the sequence wave form up to the fire four terms. The figure is shown below: Solution: Let d 0 x y f x defined as f x 0 x 2 Then by Fourier series expansion, with period 2 n x n x f x a0 an cos bn sin L L n1 n1 where L a0 an 1 2 2 0 f x dx 1 d d dx 0 dx 0 2 2 0 2 1 2L n x 1 f x cos dx d cos xdx 0 cos xdx 0 0 0 L L 2 2nd 1 bn d sin xdx 0 sin xdx 0 0 n is odd n is even Then y f x d 2d 1 1 1 sin x sin3 x sin5 x ... 2 1 3 5 206 Example 2: Find the first three terms of the Fourier series for the complete square waveform shown in the following figure: Solution: Let y f d / 2 0 t d / 2 t 2 t Notice that the square wave form is skew – symmetric. Then the Fourier series is to comprise of only sine terms. Thus, a0 0 , an 0 n bn 1 2 0 bn f t sin t d t 2 d 1 d sin t d t sin t d t 2 0 2 2d d n bn 1 1 n n 0 y f t odd even 2d 1 1 1 sin t sin3 t sin5 t ... 1 3 5 207 Problems 1. Determine whether the following functions are piecewise continuous on 1,5 : Ans a) x2 x0 f x 4 0 x 2 x x0 yes b) 1/ x 2 2 f x 2 5x 1 no c) 1 f x x2 d) f x x2 x2 lim f x 2 xx 2 no lim f x 2 xx 2 1 x2 yes, 2. Which of the following functions are piecewise smooth on a) x3 x0 f x sin x 0 x 1 x2 5x x 1 yes b) x e f x x yes x 1 x 1 c) f x ln x d) 2 x 1 f x 1/ 3 x 1 2,3 ? no limln x 0 xx 0 x 1 x 1 208 1 2/3 0 xx 3 x 1 0 no lim 3. State whether each of the following products is odd, even or neither (a) x sin x 2 (f). 2x 3 sin 4x (b) x cos x 3 (g). sin3x cos3x (c) cos2 x cos3x (h). x3e x (d) x sin x (i). x (e) 3sin x cos4 x (j). 1 cosh x x2 4 4 sin 2 x 4. Find the smallest positive period of (a) cos x,sin x,cos2 x,sin 2 x,cos x,sin x,cos2 x,sin 2 x (b) cos nx,sin nx,cos 2k x ,sin 2k x ,cos 2knx ,sin 2knx 5. Find the Fourier series of the function f x , which is a assumed to have the period 2 , and plot a accurate graphs of the first three partial sum. 1 for x 0 f x 1 for 0 x (a) f x x; x (d) (b) f x x2 ; x (e) / 4 for x 0 f x and also f f 0 f 0 / 4 for 0 x (c) f x x3; x (f) 1 for x 0 f x 1 for 0 x 1 for x / 2 (g) f x 0 for / 2 x / 2 1 for /2 x 209 6. Given that f x x x for x , find the Fourier expression for 2 f x . Deduce that 2 6 1 1 1 1 ... 22 32 42 7. State whether the given function is even or odd. Find the Fourier series. Sketch the function and some partial sums. 8. a) k if / 2 x / 2 f x 0 if / 2 x 3 / 2 b) x if / 2 x / 2 f x x if / 2 x 3 / 2 c) f x x2 / 2 x Find the Fourier cosines as well as the Fourier sine series. (a) Find a Fourier sine series for Ans: f x 1 on 0,1 . 1 1 sin n x n 2 1 n n1 (b) Find a Fourier sine series for Ans: 6 n1 f x x on 0,3 . 1n sin n x n 3 (c) Find a Fourier cosine series for f x x on 2 0, . 1 2 1 cos nx Ans: 4 2 3 n1 n n 0 x 2 on 0,3 2 x 2 (d) Find a Fourier cosine series for f x 2 4 1 2n n x Ans: sin cos 3 n1 n 3 3 (e) Find a Fourier cosine series for Ans: f x 1 on 0,7 . 1 210 x x 1 on 0,2 2 x 1 (f) Find a Fourier sine series for f x Ans: n n1 9. 4 2 2 sin n 2 n 4 n x cos cos n sin 2 n 2 n 3 Find the first four terms of the Fourier series for the saw – tooth wave form as shown below: Ans: 10. y f t 2L 1 1 1 1 sin t sin 2 t sin3 t sin 4 t ... 1 2 3 4 Determine the first four terms the Fourier series for the triangular wave form in the following figure; 211 2d 0 t / 2 t 2d Hint: f x t 2d / 2 t 3 / 2 2d 3 / 2 t 2 t 4d Ans: y f t 8d 1 1 1 1 sin t 2 sin3t 2 sin5t 2 sin7t ... 2 2 1 3 5 7 212 UNIT EIGHT TOTAL DIFFERENTIAL EQUATIONS 4.0 Recall that a function f x, y , its total differentials, df , is defined as: df f f dx dy ---------------------(٠) x y This shows that the family of curves f x, y c satisfies the differential equation df 0 . f f dx dy 0 ----------------------(٠٠) x y So if there exists a function f x, y such that: M x, y f y and N x, y x y then M x, y dx N x, y dy is called an exact differential, and the equation: M x, y dx N x, y dy 0 is said to be an exact equation, whose solution is the family f x, y c . The general form of Total Differential Equation is given as: Px1, x2 ,..., xn dx1 Qx1, x2 ,...xn dx2 ... S x1, x2 ,..xn dxn 0 ..(4.1) Examples: 1. 3x 2. 3xz 2 y dx xdy x 2 dz 0 3. ydx dy dz 0 y e x z dx 2 x 3 y sin z dy y cos z e x dz 0 2 2 It may be verified readily that example 1 is the exact differential of: f x, y, z x 3 y 2 e x z y sin z c , c is an arbitrary constant. Such an equation is called exact. Example 2 is not exact, but the use of x as an integrating factor yields: 3x z 2xydx x dy x dz 0 2 2 3 213 which is the exact differential of x z x y c . Equations 1 and 2 are called 3 2 integrable. Example 3 is not integrable; that is, no primitive (or general solution). 4.1 CONDITION OF INTEGRABILITY AND EXACTNESS The condition of integrability of the total differential equation Px, y, z dx Qx, y, z dy Rx, y, z dz 0 …(4.2) is Q R P Q R P 0 , identically….(4.3) P Q R z y x z y x And the conditions for exactness are: P Q Q R R P , , …………(4.4) y Dx z y x z Example 4.1 1. The equation 3xz 2 y dx xdy x dz 0 2 P 3xz 2 y , R x2 , P Q Q P 3x ; Q x , 1, 0 ; and 2, z x z y R R 2x , 0 .Then from (4.3),we have: x y 3xz 2 y 0 0 x2 x 3x x 2 2 1 0 x 2 x 2 0 then the equation is integrable. 2. The equation P y, ydx dy dz 0 Q Q P P R R 0 ; R 1, 0. 1, 0 ; Q 1, x z x y y z Then from (4.3),we have: y0 0 10 0 11 0 1 0 , the equation is not integrable 214 3. The equation 3x y e x z dx 2 x 3 y sin z dy y cos z e x dz 0 2 2 P 3x 2 y 2 e x z , P P e x 6x2 y , z y Q 2 x 3 y sin z , Q Q 6x 2 y , cos z x z R y cos z e x , R R e x , cos z x z Hence the equation is exact. 4.2 AN APPROACH TO SOLVE AN INTEGRABLE TOTAL DIFFERENTIAL EQUATION (in three variables) I) If the equation is exact, the solution is evident after, at most, a regrouping of terms. II) If the equation not exact, it may be possible to find an integrating factor (note: integrating factor would be given to help you solve). III) If the equation is homogeneous, one variable, say z , can be separated from the others by the transformation x uz , y vz . IV) If no integrating factor can be found, consider one of the variables, say z , as a constant. Integrate the resulting equation, denoting the arbitrary constant of integration by z , a function of z . Take the total differential of the integral just obtained and compare the coefficients of its differentials with those of the given differential equation, thus determining z . Example 4.2 1. Solve x y dx xdy zdz 0 Solution: Here P x y , Q x , R z 215 P Q Q P R R 1 0 0 , , , the equation is x x z z y x Then exact. Upon regrouping thus xdx xdy ydx zdz 0 xdx xdy ydx zdz k 12 x 2 xy 12 z 2 k x 2 2 xy z 2 c 2. Solve y dx zdy ydz 0 . 2 Solution: Here P y2 , R y, Q Q P P 0, 1, 2 y, 0 ; Q z , x z y z R R 0, 1 : x y Then Q R R Q R P P Q R x z z y y X y 2 1 1 z0 0 y2 y 0 0 and the equation is integrable. The integrable factor equation to dx x ydz zdy 0 , whose solution is: y2 z c. y 216 1 reduce the y2 3. Solve the homogeneous equation 2 y z dx x z dy 2 y x z dz 0 . Solution: The equation is integrable since: 2 y z 1 2 x z 1 2 2 y x z 2 1 0 The transformation x uz, y vz reduces the given equation to: 2zv 1udz zdu zu 1vdz zdv z2v u 1dz 0 Dividing by z and rearranging, we have: 2zv 1du zu 1dv uv u v 1dz 0 or, dividing by zuv u v 1 zu 1v 1 , 2du dv dz 0 u 1 v 1 z Then 2 lnu 1 lnv 1 ln z ln k zu 12 k v 1 x z 2 k y z y z cx z 2 . Exercise 4.1 (Solve the following) 2. 2x 1dx x dy x tan zdz 0 I.F. 1/ x 2x z zdx 2x yzdy xz xdz 0 I.F. 1/ x z 3. yzdx z 2 dy xydz 0 4. 2 y z dx 2x z dy x 2 y dz 0 1. 3 3 4 2 2 2 2 217 UNIT NINE PARTIAL DIFFERENTIAL EQUATION 5.0 NOTATION AND TERMINOLOGY Let u denote a function of several independent variables; say, u u x, y, z, t . At x, y, z, t , the partial derivative of u with respect to x is defined by u x h, y, z, t u x, y, z, t u lim ……..(5.0) x h0 h provided the limits exists. We will use the following subscript notation: u u ux uy x y 5.1 2u uxx x2 2u u yy y 2 2u 2u u yx or uxy … xy yx Definition A partial differential equation (abbreviated PDE) is an equation involving one or more partial derivatives of an unknown function of several independent variables. The order of a PDE is the order of the highest derivative that appears in the equation. The partial differential equation be linear if the function F x, y, z, t;u, ux , u y , uz , ut , uxx , uxy ,... 0 is said to F is algebraically linear in each of the variables u, ux , u y ,..., and if the coefficients of u and its derivatives are functions only of the independent variables. An equation that is not linear is said to be nonlinear; a nonlinear equation is quasilinear if it is linear in the highest – order derivatives. Example 5.1 z z y z is of order one. x y 1. x 2. 2 z 2 z 2 z 3 0 is of order two. xy y 2 x 2 3. 2u 2 2u c t 2 x2 One – dimensional wave equation 218 4. 2u 2 2u c t x2 One – dimensional heat equation 5. 2u 2u 0 x2 y 2 Two – dimensional Laplace equation 6. 2u 2u f x, y x2 y 2 Two – dimensional Poisson equation 7. 2u 2 2u 2u c 2 2 t 2 y x Two – dimensional wave equation 8. 2u 2u 2u 0 x2 y 2 z 2 Three – dimensional Laplace equation Their solution is in the form: 5.2 u f x, y, z, t ,... SOLUTION TO A FIRST ORDER PARTIAL DIFFERENTIAL EQUATION A linear partial differential equation of order one, involving a dependent variable z and two independent variables x and y , is of the form: P where z z Q R ……………………(5.1) x y P, Q, R are functions of x, y, z . If P 0 , or Q 0 , equation (5.1) may be solved easily. Thus, the equation z 2 x 3 y has as solution z x 2 3xy y , where is an arbitrary x function. Lagrange reduced the problem of finding the general solution of 5.1 to that of solving an auxiliary system (called the Lagrange system) of ordinary differential equations: dx dy dz ……………….(5.2) P Q R 219 by showing that: u, v 0 , arbitrary is the general solution of equation (5.1) provided u ux, y, z a and v vx, y, z b are two independent solutions of equation (5.2). Here, a ………(5.3) and b are arbitrary constants and at least one of u, v must contain z . Example 5.1 Find the general solution of x z z y 3z . x y Solution: The auxiliary system is: dx dy dz . x y z From v dx dz dx dz z , we obtain u 3 a ; and from , we obtain x 3z x y x y b. x Thus, the general solution is Of course, from z y , 0 , where is arbitrary. x3 x dy dz z , we obtain 3 c , and we may write y 3z y z z z y 0 or , 0 , 3 3 y3 x x y where and are arbitrary. However, these are all equivalent and we shall call any one of them the general solution 220 COMPLETE SOLUTIONS If u a and v b are two independent solutions of equations (5.2) and if , are arbitrary constants, u v is called a complete solution of (5.1). Thus, for the equation is the last example above () z / x 3 y / x is a complete solution. Example 5.2 Find the general solution of 2 z z 3 1 x y Solution: The auxiliary system is : From dx dy dz 2 3 1 dx dy dx dz , we have , we have x 2 z a ; and from 2 3 2 1 3x 2 y b . Thus, the general solution is: x 2 y,3x 2 y 0 . The complete solution x 2 z 3x 2 y is a two – parameter of family of planes. 5.3 TYPES OF SECOND – ORDER EQUATIONS In the linear PDE of order two in two variables, auxx 2buxy cu yy dux eu y fu g ………………..(5.3) if uxx is formally replaced by 2 , uxy by , u yy by 2 , u x then associated with (5.3) is a polynomial of degree two in P , a 2 2b c 2 d e f 221 by ,and u y by , and : The mathematical properties of the solutions of (5.3) are largely determined by the algebraic properties of the polynomial P , . P , - and along with it, the PDE (5.3) – is classified as hyperbolic, parabolic, pr elliptic according as its discriminant, b2 ac is positive, zero, or negative. Note that the type of (5.3) is determined solely by its principal part (the terms involving the highest – order derivatives of u ) and that the type will generally change with position in the xy plane unless a, b, and c are constants. Example 5.2 1. The PDE 3uxx 2uxy 5uyy xuy 0 is elliptic, since b2 ac 12 3 5 14 0 2. The Tricomi equation for transonic flow, uxx yuyy 0 , has b2 ac 02 1 y y Thus, the equation is elliptic for hyperbolic for y 0 , parabolic for y 0 , and y 0. The general linear PDE of order two in n variables has the form: n a u i , j 1 If ij xi x j n biuxi cu d ……………………….(5.4) i 1 uxi x j ux j xi , then the principal part of (5.4) can always be arranged so that aij a ji ; thus, the n n matrix A aij can be assumed symmetric. In linear algebra it is shown that every real, symmetric n n matrix has n real eigenvalues. These eigenvalues are the (possibly repeated) zeros of an n th – degree polynomial in , det A I 0 , where I is the n n identity matrix. Let P denote the number of positive eigenvalues, and Z the number of zero eigenvalues (i.e., the multiplicity of the eigenvalues zero), of the matrix A . Then (5.4) is: hyperbolic if Z 0 and P 1 or Z 0 and P n 1 222 parabolic if Z 0 (equivalently, if det A 0 ) elliptic if Z 0 and P n or Z 0 and P 0 ultrahyperbolic if Z 0 and 1 P n 1 if any of the aij is nonconstant, the type of (5.4) can vary with position. Example 5.3 For the PDE 3ux1x1 ux2 x2 4ux2 x3 4ux3x3 0 , 0 0 3 0 0 3 A 0 1 2 and det 0 1 2 3 5 0 0 2 4 0 2 4 Because 0 is an eigenvalues, the PDE is parabolic (throughout x1x2 x3 space). Exercise 5.1 1. Classify according to type: a) uxx 2 yuxy xu yy ux u 0 b) 2xyuxy xu y yux 0 c) uxx uxy 5uyx uyy 2uyz uzz 0 2. Describe the regions where the equation is hyperbolic, parabolic, elliptic: a) uxx uxy 2u yy 0 b) uxx 2uxy u yy 0 c) 2uxx 4uxy 3uyy 5u 0 d) uxx 2 xuxy u yy cos xy ux u e) yuxx 2uxy exuyy u 3 f) exyuxx sinh x uyy u 0 g) xuxx 2xyuxy yu yy 0 h) xuxx 2xyuxy yu yy 0 223 SOLUTIONS TO SECOND ORDER LINEAR PARTIAL DIFFERENTIAL EQUATION. Example 5.3 u 2u 2 sin t . x 0 u cos2 t 12 x t 1 Solve given that , and x x2 Solution u 12 x2 t 1 dx 4 x3 t 1 t x u x4 t 1 x t t Using initial conditions t sin t hence x 0; u sin t at u cos2t x t cos2t u x4 t 1 x sin t cos2t . Exercise 2u sin x y , given that Solve the equation xy y 0, u 2 1; x 0, u y 1 x Initial Conditions and Boundary Conditions As with any differential equation, the arbitrary constant or arbitrary functions in any particular case are determine from initial conditions or, more generally, the boundary conditions since they do not always refer to zero values of independent variables. 224 5.4 SEPARATION OF VARIABLES Solving second – order PDE with two independent variables we assume a trial solution of the form u x, t x t (or simply u )where; x is a function of x only. t is a function of t only. Therefore u 2u 2 ; x x u 2u 2 t t 2u 2 2u 1 c Hence; the Wave equation can be written as . c2 t 2 x2 the Heat equation 2u 2 2u 1 c can be written as . 2 t c2 x Y 2u 2u the Laplace equation can be written as . 0 Y x2 y 2 5.5 SOLUTION OF THE WAVE EQUATION u f x, t P u x, t 0 l x 2u 1 2u The wave equation has a solution u x, t . x2 c2 t 2 Boundary conditions: a) The string is fixed at both ends, i.e. at x 0 and x l for all values of time t . Therefore u x, t becomes u 0, t 0 for all values of t 0 u l, t 0 225 Initial conditions: b) If the initial deflection of P at t 0 is denoted by f x , then u x,0 f x . c) Let the initial velocity of u P be g x , then g x t t 0 2u 1 2u 1 The wave equation becomes . c2 x2 c2 t 2 Let 1 k also 2 k c k 0 and c2k 0 For oscillatory motion let k p i.e. k 0 . 2 p 2 0 Acos px B sin px and also c p 0 C cos cpt Dsin cpt 2 2 so our suggested solution u now becomes u x, t Acos px B sin px C cos cpt D sin cpt and, if we put cp p / c u x, t Acos x B sin x C cos t D sin t c c where A, B, C, D are arbitrary constants. Now, using the boundary conditions u 0 and x 0 0 AC cos t D sin t t A 0 Therefore, u0 u x, t B sin x C cos t D sin t c x l x 0 B sin Now, B 0 or l c C cos t D sin t u x, t would be identically zero. 226 sin l c 0 l c n nc for n 1,2,3,... l As we can see, there is an infinite set of values of value gives a particular solution for and each separate u x, t . The values of are called the eigenvalues and each corresponding solution the eigenfunction. Eigenvalues Eigenfunctions nc l u x, t B sin x C cos t D sin t n 1 1 c l u1 sin 2 2 2c l u2 sin 2 x 2c t 2c t D2 sin C2 cos l l l 3 3 3c l u3 sin 3 x 3c t 3c t D3 sin C3 cos l l l r r rc l ur sin r x rc t rc t Dr sin Cr cos l l l c x c t c t D1 sin C1 cos l l l Note BC Cn and BD Dn , where C1, C2 , C3 ,.... and D1, D2 , D3 ,.... are arbitrary constants. If u1, u2 , u3 ,.... are solutions, then the linear combination is also a solution, which is a more general solution, i.e. u u1 u2 u3 So rc t rc t r x u x, t ur sin Dr sin Cr cos l l l r 1 r 1 To solve for the arbitraries Cr , Dr we use the initial conditions which we have not yet taken into account. 227 Example5.2 A stretched string of length 20cm and is set oscillating by displacing its midpoint a distance 1cm from its rest position and releasing it with zero initial velocity. Solve 2u 1 2u the wave equation 2 2 where c2 1 to determine the resulting motion, 2 x c t u x, t . Solution: 2 1 20 10 u x 10 u 2 x 10 From data given the boundary conditions are u 0, t 0; u 20, t 0 fixed endpoints. x 0 x 10 10 u x,0 f x 20 x 10 x 20 10 u t (zero initial velocity) t 0 Since c 1 , then we have Which implies p 0 and 2 p2 0 The respective solutions are Acos px B sin px and C cos pt Dsin pt u x, t Acos px B sin px C cos pt D sin pt u x, t Acos x B sin x C cos t D sin t where cp but c 1 228 u 0, t 0, x 0 0 AC cos t D sin t A 0 x, t B sin x C cos t D sin t u 20, t 0, x 20 0 B sin 20 C cos t D sin t Therefore, B 0 or u would be identically zero sin 20 0 20 n Hence, u x, t B sin n 20 n n n x C cos t D sin t 20 20 20 Follow the table below: Eigenvalues Eigenfunctions n 20 u x, t B sin xC cos t D sin t n 1 1 2 2 2 20 u2 sin 2 x 2 t 2 t D2 sin C2 cos 20 20 20 3 3 3 20 u3 sin 3 x 3 t 3 t D3 sin C3 cos 20 20 20 r r r 20 ur sin r x r t r t Dr sin Cr cos 20 20 20 u1 sin 20 x t t C1 cos D1 sin 20 20 20 Note BC Cn and BD Dn , where C1, C2 , C3 ,.... and D1, D2 , D3 ,.... are arbitrary constants. u u1 u2 u3 r t r t r x u x, t ur sin Dr sin Cr cos 20 20 20 r 1 r 1 229 To find Cr and Dr we use the rest of the conditions i.e.; x 0 x 10 10 u x,0 f x 20 x 10 x 20 10 u x,0 Cr sin r 1 Then Cr 2 mean value of f x sin Cr 10 0 10Cr 20 20 x x r x r x sin dx sin dx 10 10 10 20 20 20 r 40 r 20 r 40 cos 2 2 sin cos 2 2 sin r r 2 r 2 r 2 r 4 For r=1,2,3,... Cr u x, t sin r 1 and for the condition r x between x 0 and x 20 20 2 20 r x f x sin dx 20 0 20 10Cr Then r x 20 r 2 2 sin r 2 r x 4 r r t r t Dr sin 2 2 sin cos 20 r 2 20 20 u u t 0; 0 (zero initial velocity), which is t t t 0 u x, t r x 4 r r r t r r t sin sin sin D cos r t 20 r 2 2 2 20 20 20 20 r 1 Therefore, at t 0 , 0 sin r 1 So finally we have u x, t r x r D 20 r 20 4 2 1 r 2 sin r 1 Dr 0 r x r r t sin cos 20 2 20 230 5.5 HEAT AND LAPLACE EQUATION The Heat and Laplace equations look slightly different from the wave equation, but the method of solution is very much along the same lines. 1. For Heat equations; c2uxx ut 1 c2 p2 0 given = A cos px B sin px p2c2 0 given Ce p c t 2 2 u x, t Acos px B sin px Ce p c t 2 2 pc 2. 2 u x, t AC cos x BC sin x e t c c For Laplace’s equation: uxx u yy 0 Y Y p2 0 given = A cos px B sin px Y p2Y 0 given Y C cosh py D sinh py, which we can express as: E sinh p y u x, y Acos px B sin px E sinh p y Exercise 5.2 1. A bar of length 2m is fully insulated along its sides. It is initially at a uniform temperature of 10C and at t 0 the ends are plunged into ice and maintained at a temperature of 0C . Determine an expression for the temperature at a point P a distance x from one end at any subsequent time t seconds after t 0 . 231 2. 2u 2u Determine a solution u x, y of the Laplace equation 0 x2 2 y subject to the following boundary conditions: u 0 when x 0; u 0 when x ; u 0 when y ; u 3 when y 0 . SOLUTIONS TO SECOND ORDER LINEAR PARTIAL DIFFERENTIAL EQUATION. Example 5.3 u 2u sin t . 12 x2 t 1 given that x 0 , u cos2t and Solve 2 x x Solution u 12 x2 t 1 dx 4 x3 t 1 t x u x4 t 1 x t t Using initial conditions x 0; u sin t at u cos2t x 232 t sin t hence t cos2t u x4 t 1 x sin t cos2t . Exercise Solve the equation y 0, 2u sin x y , given that xy u 2 1; x 0, u y 1 x Initial Conditions and Boundary Conditions As with any differential equation, the arbitrary constant or arbitrary functions in any particular case are determine from initial conditions or, more generally, the boundary conditions since they do not always refer to zero values of independent variables. 5.4 SEPARATION OF VARIABLES Solving second – order PDE with two independent variables we assume a trial solution of the form u x, t x t (or simply u )where; x is a function of x only. t is a function of t only. u 2u 2 ; Therefore x x Hence; the Wave equation the Heat equation u 2u 2 t t 2u 2 2u 1 c can be written as . c2 t 2 x2 2u 2 2u 1 c can be written as . t c2 x2 Y 2u 2u . the Laplace equation 2 0 can be written as 2 Y x y 5.5 SOLUTION OF THE WAVE EQUATION 233 u f x, t P u x, t 0 l x 2u 1 2u The wave equation has a solution u x, t . x2 c2 t 2 Boundary conditions: d) The string is fixed at both ends, i.e. at x 0 and x l for all values of time t . Therefore u x, t becomes u 0, t 0 for all values of t 0 u l, t 0 Initial conditions: e) If the initial deflection of P at t 0 is denoted by f x , then u x,0 f x . f) Let the initial velocity of u P be g x , then g x t t 0 2u 1 2u 1 The wave equation becomes . c2 x2 c2 t 2 Let 1 k also 2 k c k 0 and c2k 0 For oscillatory motion let k p i.e. k 0 . 2 p 2 0 and also c p 0 2 2 Acos px B sin px C cos cpt Dsin cpt 234 so our suggested solution u now becomes u x, t Acos px B sin px C cos cpt D sin cpt and, if we put cp p / c u x, t Acos x B sin x C cos t D sin t c c where A, B, C, D are arbitrary constants. Now, using the boundary conditions u 0 and x 0 0 AC cos t D sin t t A 0 u x, t B sin x C cos t D sin t c Therefore, x l x 0 B sin u0 Now, B 0 or sin l c l c C cos t D sin t u x, t would be identically zero. 0 l c n nc for n 1,2,3,... l As we can see, there is an infinite set of values of value gives a particular solution for and each separate u x, t . The values of are called the eigenvalues and each corresponding solution the eigenfunction. Eigenvalues Eigenfunctions nc l u x, t B sin x C cos t D sin t n 1 1 c l u1 sin 2 2 2c l u2 sin 2 x 2c t 2c t D2 sin C2 cos l l l 3 3 3c l u3 sin 3 x 3c t 3c t D3 sin C3 cos l l l c x c t c t D1 sin C1 cos l l l 235 r r rc l ur sin r x rc t rc t Dr sin Cr cos l l l Note BC Cn and BD Dn , where C1, C2 , C3 ,.... and D1, D2 , D3 ,.... are arbitrary constants. If u1, u2 , u3 ,.... are solutions, then the linear combination is also a solution, which is a more general solution, i.e. u u1 u2 u3 So rc t rc t r x u x, t ur sin Dr sin Cr cos l l l r 1 r 1 To solve for the arbitraries Cr , Dr we use the initial conditions which we have not yet taken into account. Example5.2 A stretched string of length 20cm and is set oscillating by displacing its midpoint a distance 1cm from its rest position and releasing it with zero initial velocity. Solve 2u 1 2u 2 2 where c2 1 to determine the resulting motion, the wave equation 2 x c t u x, t . Solution: 236 2 1 20 10 u x 10 u 2 x 10 From data given the boundary conditions are u 0, t 0; u 20, t 0 fixed endpoints. x 0 x 10 10 u x,0 f x 20 x 10 x 20 10 u t (zero initial velocity) t 0 Since c 1 , then we have Which implies p 0 and 2 p2 0 The respective solutions are Acos px B sin px and C cos pt Dsin pt u x, t Acos px B sin px C cos pt D sin pt u x, t Acos x B sin x C cos t D sin t where cp but c 1 u 0, t 0, x 0 0 AC cos t D sin t A 0 x, t B sin x C cos t D sin t u 20, t 0, x 20 0 B sin 20 C cos t D sin t Therefore, B 0 or u would be identically zero sin 20 0 20 n Hence, u x, t B sin n n n x C cos t D sin t 20 20 20 237 n 20 Follow the table below: Eigenvalues Eigenfunctions n 20 u x, t B sin xC cos t D sin t n 1 1 2 2 2 20 u2 sin 2 x 2 t 2 t D2 sin C2 cos 20 20 20 3 3 3 20 u3 sin 3 x 3 t 3 t D3 sin C3 cos 20 20 20 r r r 20 ur sin r x r t r t Dr sin Cr cos 20 20 20 20 u1 sin x t t C1 cos D1 sin 20 20 20 Note BC Cn and BD Dn , where C1, C2 , C3 ,.... and D1, D2 , D3 ,.... are arbitrary constants. u u1 u2 u3 r t r t r x u x, t ur sin Dr sin Cr cos 20 20 20 r 1 r 1 To find Cr and Dr we use the rest of the conditions i.e.; 238 x 0 x 10 10 u x,0 f x 20 x 10 x 20 10 u x,0 Cr sin r 1 Then Cr 2 mean value of f x sin Cr r x between x 0 and x 20 20 2 20 r x f x sin dx 20 0 20 10Cr 10 0 Then r x 20 10Cr 20 20 x x r x r x sin dx sin dx 10 10 10 20 20 20 r 40 r 20 r 40 cos 2 2 sin cos 2 2 sin r r 2 r 2 r 2 r For r=1,2,3,... Cr u x, t sin r 1 and for the condition 4 r 2 sin 2 r 2 r x 4 r r t r t Dr sin 2 2 sin cos 20 r 2 20 20 u u t (zero initial velocity), which is t 0; t 0 t 0 u x, t r x 4 r r r t r r t sin 2 2 sin sin Dr cos t 20 r 2 20 20 20 20 r 1 Therefore, at t 0 , 0 sin r 1 So finally we have u x, t 4 r x r D 20 r 20 1 sin 2 r2 r 1 Dr 0 r x r r t sin cos 20 2 20 239 5.5 HEAT AND LAPLACE EQUATION The Heat and Laplace equations look slightly different from the wave equation, but the method of solution is very much along the same lines. 3. For Heat equations; c2uxx ut 1 c2 p2 0 given = A cos px B sin px p2c2 0 given Ce p c t 2 2 u x, t Acos px B sin px Ce p c t 2 2 pc 4. 2 u x, t AC cos x BC sin x e t c c For Laplace’s equation: uxx u yy 0 Y Y p2 0 given = A cos px B sin px Y p2Y 0 given Y C cosh py D sinh py, which we can express as: E sinh p y u x, y Acos px B sin px E sinh p y Exercise 5.2 3. A bar of length 2m is fully insulated along its sides. It is initially at a uniform temperature of 10C and at t 0 the ends are plunged into ice and maintained at a temperature of 0C . Determine an expression for the temperature at a point P a distance x from one end at any subsequent time t seconds after t 0 . 240 4. 2u 2u Determine a solution u x, y of the Laplace equation 0 x2 2 y subject to the following boundary conditions: u 0 when x 0; u 0 when x ; u 0 when y ; u 3 when y 0 . 241