Solutions of Ordinary Differential Equations Consider the n-th order ODE F(x, y, y!, y!!,..., y( n ) ) = 0 (1). Definition: Let g(x) be a real-valued function defined on a interval I, having the n-th derivative for all x in I. g(x) is called an explicit solution of the equation (1) on the interval I, if: 1) F(x,g(x), g!(x), g!!(x),...,g( n ) (x)) is defined for all x in I 2) F(x,g(x), g!(x), g!!(x),...,g( n ) (x)) = 0 for all x in I Example: 1) Verify that the function g(x) = e2x is an explicit solution of the equation F(x,y,y’,y’’) = y’’ + y’ – 6y = 0 We have g’(x) = 2e2x and g’’(x) = 4e2x, when we substitute in the equation, we get F(x, g(x), g’(x), g’’(x))= 4e2x + 2e2x – 6 e2x = 0 and defined for all x real. 2 x3 2) Verify that the function h(x) = defined on the interval (0, 3) is an explicit solution of the equation F(x,y,y’) = 3xy’ – 2y = 0. 1 2 ! Since g’(x) = x 3 is defined on the interval (0, 3) and 3 1 2 2 2 2 ! F(x, g(x), g’(x)) = 3x x 3 ! 2x 3 = 2x 3 ! 2x 3 = 0 and defined for all x in (0, 3) 3 Definition: A relation H(x,y) = 0 is called an implicit solution of the ODE (1) if this relation produces at least one real-valued function g(x) defined on the interval I, such that g(x) is an explicit solution of (1) on I. Example: 1) The relation x2 + y2 – 4 = 0 is an implicit solution of the equation F(x,y,y’) = x + yy’= 0 on the interval (-2, 2). The relation produces the functions h(x) = y = 4 ! x 2 and also the function k(x) = y = - 4 ! x 2 both defined on (-2, 2). "2x Since h!(x) = , then 2 2 4"x !x 2 F(x,h(x),h’(x)) = x + 4 ! x 2 = 0 and defined on (-2, 2). 4 ! x2 2) The relation x2 + y2 + 4 = 0 is NOT an implicit solution of the equation F(x,y,y’) = x + yy’= 0 on the interval (-2,2). The relation does not produce a real-valued function on the interval, solving for y we get p(x) = y = !4 ! x 2 and this function is undefined since - 4 – x2 < 0. If we differentiate the relation implicitly, we get 2x + 2yy’ = 0 or x + yy’ = 0, that is the equation we want to solve. Then, we say that the relation x2 + y2 + 4 = 0 is a formal solution of the differential equation. Remark: A first order ODE can be given by the expression dy = f(x, y) dx Geometric interpretation of the first order ODE dy A first order ODE = f(x, y) associates with each point (x0, y0) in a region D ! ! 2 a dx dy direction m = = f(x 0 , y 0 ) . The direction at each point (x0, y0) is the slope of the dx tangent line to a curve, with equation g(x,y) = c, passing through the point. The region D with the direction at each point creates what is called a direction field. Solving the differential equation means to find curves whose tangent lines at the point dy (x0, y0) has slope m = = f(x 0 , y 0 ) . dx Example: dy The ODE = 2x creates a directional field in ! 2 . Every curve y = x2 + c where c is dx an arbitrary constant, has a tangent line at the point (x,y) with slope m = 2x. Remark: We saw in the above example that functions of the form hc(x) = x2 + c where c is any dy constant, are solutions of the ODE = 2x . The constant c is called a parameter. dx Then, hc(x) = x2 + c is a one-parameter family of solutions of the ODE. Every first order ODE has a one-parameter family of solutions. Initial Value Problem Let’s consider the following problem: Find a function k(x) that is a solution of the ODE dy = 2x , such that k(1) = 4. dx This means that: 1- k(x) must satisfy the ODE, so k’(x) = 2x for all x 2- k(1) = 4 This is called an initial value problem (I.V.P.) and it is denoted by: ! dy # = 2x " dx #$ y(1) = 4 dy = 2x has a one-parameter family of solutions hc(x) = x2 + c where c dx is an arbitrary constant, imposing the condition x =1 then y = 1, we get 4 = (1)2 + c or c = 3. Therefore, k(x) = x2 + 3 satisfies the I.V.P.. Since the ODE Remark: For any first order ODE, we denote an I.V.P. by ! dy # = f(x, y) " dx #$ y(x 0 ) = y 0 The I.V.P. can be extended to ODE of higher order. Consider the problem " d2y $ 2 +y=0 $$ dx # y(0) = !1 $ y '(0) = 0 $ $% We have to find a solution h(x) such that satisfies the ODE and moreover h(0) = -1 and h’(0) = 0. For ODE of order 2 or higher there is another type of problem called boundary-value problem and it is given by " d2y $ 2 +y=0 $$ dx # y(0) = 1 $ y(!) = 0 $ $% The conditions relate to the two different values of x, 1 and π. Theorem: Basic Existence and Uniqueness Theorem Given the I.V.P. ! dy # = f(x, y) " dx #$ y(x 0 ) = y 0 !f (x, y) are continuous in a rectangle If the function f(x,y) and the partial derivative !y R centered at (x0,y0), R = {(x,y): x0- h < x < x0 + h, y0- k < y < y0 + k} Then, there is a function p(x) defined on the interval (x0 – h, x0 + h) that is the unique solution of the I.V.P. Examples: 1) Consider the I.V.P. x " dy $ =! y # dx $ y(3) = 4 % The relation x2 + y2 = c2 is a one-parameter family of solution of the ODE dy x =! . dx y The condition y(3) = 4 implies 32 + 42 = 25 = c2 Solving for y we get to functions: h(x) = 25 ! x 2 and k(x) = - 25 ! x 2 both defined on the interval [-5,5]. Since h(3) = 4 and k(3) = -4, the only solution of the I.V.P. is h(x) = 25 ! x 2 . x !f x Notice that f(x,y) = ! and = 2 are continuous on a rectangle centered at (3,4) and y !y y does not include the x-axis. 2) Consider the I.V.P. x " dy $ =! y # dx $ y(1) = 0 % x "f x Since f(x, y) = ! !!and!! = are NOT continuous along the x-axis, then we cannot y "y y !f create a rectangle centered at (1,0) where f and are continuous. The theorem fails and !y the I.V.P. has two solutions h(x) = 1 ! x 2 and k(x) = - 1 ! x 2 3) Consider the I.V.P. ! dy # =2 y " dx #$ y(0) = 0 !f 2 is NOT defined at (0,0), then we cannot create a rectangle centered at = !y 2 y !f (0,0) where both f and are continuous, therefore the theorem fails and the I.V.P. has !y two different solutions, h(x) = x2 and k(x) = 0. Since