Outline 1 Motivation 2 Basic Ideas and Terminology 3 Solutions of Differential Equations 4 Initial-Value Problem 5 Separable Differential Equations 6 Some Simple Population Models 7 First-Order Linear Differential Equations 8 Modeling Problems Using First-order Linear Differential Equations 9 Change of Variables 10 Exact Differential Equations 11 Numerical Solution to First-order Differential Equations MATH2030 Differential Equations (AY 2023-24) 104 / 177 Change of Variables In this section, we consider two further types of differential equations which can be solved by using a change of variables to reduce them to one of the types we know how to solve. Definition A function f (x, y ) is said to be homogeneous of degree zero if f (tx, ty ) = f (x, y ) for all positive values of t for which (tx, ty ) is in the domain of f . Equivalently, f is homogeneous of degree zero if it is invariant under a re-scaling of the variables x and y . MATH2030 Differential Equations (AY 2023-24) 105 / 177 Examples 1 The simplest non-constant functions which are homogeneous of y x degree zero are f (x, y ) = and f (x, y ) = . x y y Indeed, for say f (x, y ) = we have x f (tx, ty ) = 2 If f (x, y ) = ty y = = f (x, y ). tx x x2 − y2 , then 2xy + y 2 f (tx, ty ) = t 2 (x 2 − y 2 ) = f (x, y ), t 2 (2xy + y 2 ) so that f is homogeneous of degree zero. MATH2030 Differential Equations (AY 2023-24) 106 / 177 Note that in the second example, dividing both the numerator and denominator by x 2 , we have x2 − y2 = f (x, y ) = 2xy + y 2 y2 x2 − 2 x x2 2 2xy + yx 2 x2 = 2 1− y x y 2 x 2 . + yx This means that f can be considered to depend on the single variable y V = . x MATH2030 Differential Equations (AY 2023-24) 107 / 177 Actually, we have the following result that establishes the basic property of all functions that are homogeneous of degree zero. Theorem A function f (x, y ) is homogeneous of degree zero if and only if it y x depends on either or only. x y MATH2030 Differential Equations (AY 2023-24) 108 / 177 Homogeneous first-order differential equations We now consider differential equations that satisfy the following definition. Definition If f (x, y ) is homogeneous of degree zero, then the differential equation dy = f (x, y ) dx (1) is called a homogeneous first-order differential equation. MATH2030 Differential Equations (AY 2023-24) 109 / 177 In general, a homogeneous first-order differential equation cannot be solved directly. However, we can solve it, by the theorem above, using a change of variables: instead of using variables x and y , we can use the variables y x and V , where V = , that is y = xV (x). x MATH2030 Differential Equations (AY 2023-24) 110 / 177 Theorem The change of variables y = xV (x) reduces a homogeneous dy first-order differential equation = f (x, y ) to the separable equation dx 1 1 dV = dx, F (V ) − V x where F is an appropriate function. Proof Exrecise. MATH2030 Differential Equations (AY 2023-24) 111 / 177 Remark The separable equation said above can be integrated to obtain a relationship between V and x. Then the solution to the given y differential equation can be obtained by substituting for V in this x relationship. Examples Find the general solution to dy 4x + y = . dx x − 4y Solution Since f (x, y ) = 4x+y x−4y is homogeneous of degree zero, the given equation is a first-order homogeneous differential equation. MATH2030 Differential Equations (AY 2023-24) 112 / 177 Substituting y = xV into the equation yields 4+V d (xV ) = , dx 1 − 4V that is x dV 4+V +V = , dx 1 − 4V or equivalently, x 4(1 + V 2 ) dV = . dx 1 − 4V Separating the variables gives 1 − 4V 1 dV = dx. 2 x 4(1 + V ) MATH2030 Differential Equations (AY 2023-24) 113 / 177 We write this as 1 V 1 − dV = dx. 2 2 x 4(1 + V ) 1 + V Now we integrate both side directly. For this the following integral formulas are used: Z Z dx x dx 1 (1) = arctan x + C, and (2) = log(1 + x 2 ) + C. 2 2 2 1+x 1+x The first integral is standard: the change of variables x = tan t works well. For the second integral: putting 1 + x 2 = t we have dt = 2x dx, and hence Z Z x dx 1 dt 1 1 = = log |t| + C = log(1 + x 2 ) + C. 2 2 t 2 2 1+x MATH2030 Differential Equations (AY 2023-24) 114 / 177 Thus we obtain 1 1 arctan V − log(1 + V 2 ) = log |x| + C. 4 2 y and multiplying through by 2 yields x 2 y 1 x + y2 arctan − log = log(x 2 ) + C1 , 2 x x2 Substituting back V = where C1 =2C. 2 2 Moving log x x+y to the right-hand side yields the answer 2 y 1 arctan − log(x 2 + y 2 ) = C, 2 x where C is an arbitrary constant. MATH2030 Differential Equations (AY 2023-24) 115 / 177 Remark The solution to the example above is more easily expressed in terms of polar coordinates. x = r cos θ Indeed, note that y = r sin θ obtain a logarithmic spiral: p r = x 2 + y 2 ⇐⇒ θ = arctan y , we then x θ r = Ce 4 MATH2030 Differential Equations (AY 2023-24) 116 / 177 Bernoulli equations We now consider a special type of nonlinear differential equation that can be reduced to a linear equation by a change of variables. Definition A differential equation that can be written in the form dy + p(x)y = q(x)y α , dx (2) where α is a real constant, is called a Bernoulli equation. If α = 0 or α = 1, Equation (2) is linear, but otherwise it is nonlinear. In general, we can reduce Equation (2) to a linear equation as follows. MATH2030 Differential Equations (AY 2023-24) 117 / 177 Divide Equation (2) by y α to obtain y −α dy + y 1−α p(x) = q(x). dx (3) Make the change of variables u(x) = y 1−α to get dy du = (1 − α)y −α . dx dx Thus y −α dy 1 du = . dx 1 − α dx MATH2030 Differential Equations (AY 2023-24) 118 / 177 Now substituting into Equation (3) the expressions y 1−α = u(x) and dy 1 du y −α = , we have the linear differential equation dx 1 − α dx 1 du + p(x)u = q(x), 1 − α dx or in the standard form du + (1 − α)p(x)u = (1 − α)q(x). dx MATH2030 Differential Equations (AY 2023-24) 119 / 177 Examples Solve 2 dy 3 12y 3 + y=√ , x > 0. dx x 1 + x2 Solution The differential equation is Bernoulli equation with α = 23 . 2 Dividing both sides of the equation by y α = y 3 we obtain 2 y−3 dy 3 1 12 + y3 = √ . dx x 1 + x2 (4) We make the change of variables 1 u = y 1−α = y 3 . MATH2030 Differential Equations (AY 2023-24) 120 / 177 Then we get 1 2 dy du = y−3 . dx 3 dx Now substitute the last two expressions into Equation (4) yields 3 du 3 12 + u=√ , dx x 1 + x2 or in the standard form du 1 4 + u=√ . dx x 1 + x2 (5) This is a first-order linear differential equation, whose integrating factor is MATH2030 Differential Equations (AY 2023-24) 121 / 177 I(x) = e R 1 dx x = elog x (because x > 0) = x. So Equation (5) can be written as d 4x (xu) = √ . dx 1 + x2 Integrating this equation, we obtain p xu(x) = 4 1 + x 2 + C. Here we use the integral formula Z p x dx √ = 1 + x 2 + C, 1 + x2 which is computed as follows: putting 1 + x 2 = t implies x dx = MATH2030 Differential Equations (AY 2023-24) dt 2; 122 / 177 then Z x dx √ = 1 + x2 p √ dt √ = t + C = 1 + x 2 + C. 2 t Z Returning back to the obtained result p xu(x) = 4 1 + x 2 + C, we have 1 p 4 1 + x2 + C , x and the solution to the original differential equation is u(x) = 1 y3 = 1 p 4 1 + x2 + C , x where C is an arbitrary constant. MATH2030 Differential Equations (AY 2023-24) 123 / 177 Outline 1 Motivation 2 Basic Ideas and Terminology 3 Solutions of Differential Equations 4 Initial-Value Problem 5 Separable Differential Equations 6 Some Simple Population Models 7 First-Order Linear Differential Equations 8 Modeling Problems Using First-order Linear Differential Equations 9 Change of Variables 10 Exact Differential Equations 11 Numerical Solution to First-order Differential Equations MATH2030 Differential Equations (AY 2023-24) 124 / 177 Exact Differential Equations The technique presented in this section is for first-order differential equations written in a differential form M(x, y ) dx + N(x, y ) dy = 0. Recall from a calculus course that if u = u(x, y ) is a function of two variables x and y , then its differential du is defined by du = ∂u ∂u dx + dy . ∂x ∂y MATH2030 Differential Equations (AY 2023-24) 125 / 177 Example Solve the equation 2x sin y dx + x 2 cos y dy = 0. (1) Solution This equation is separable, but we will use a different technique to solve it. We notice that 2x sin y dx + x 2 cos y dy = d(x 2 sin y ). Then Equation (1) can be written as d(x 2 sin y ) = 0, which implies that x 2 sin y is constant, denoted by C. Thus the general solution to Equation (1) is sin y = C , x2 where C is an arbitrary constant. MATH2030 Differential Equations (AY 2023-24) 126 / 177 Obviously, in the example above, it was able to write the given differential equation in the form du(x, y ) = 0, and hence obtain its solution. However, we cannot always do it. Indeed, we see that the differential equation M(x, y ) dx + N(x, y ) dy = 0 can be written as du = 0 if and only if M= ∂u ∂x and N = ∂u ∂y for some function u. This leads to the following definition: MATH2030 Differential Equations (AY 2023-24) 127 / 177 Definition The differential equation M(x, y ) dx + N(x, y ) dy = 0 (2) is said to be exact in a domain D of the xy -plane if there exists a function u(x, y ) such that ∂u ∂u = M, = N, ∂x ∂y (3) for all (x, y ) ∈ D. Any function u satisfying (3) is called a potential function for the Equation (2). MATH2030 Differential Equations (AY 2023-24) 128 / 177 If such a function does exist, then Equation (2) can be written as M(x, y ) dx + N(x, y ) dy = 0 ∂u ∂u ⇐⇒ dx + dy = 0 ∂x ∂y ⇐⇒ du = 0. This is a reason why such a differential equation is called an exact differential equation. For the example above, the potential function is u(x, y ) = x 2 sin y . MATH2030 Differential Equations (AY 2023-24) 129 / 177 The following result shows that if a differential equation is exact and if a potential function u can be found, then we can write down its solution immediately. Theorem The general solution to an exact differential equation M(x, y ) dx + N(x, y ) dy = 0 is defined implicitly by u(x, y ) = C, (4) where u(x, y ) satisfies (3) and C is an arbitrary constant. MATH2030 Differential Equations (AY 2023-24) 130 / 177 Proof Exrecise. MATH2030 Differential Equations (AY 2023-24) 131 / 177 There naturally arise 2 questions: 1 How to verify if a given differential equation is exact? 2 How to find a potential function for an exact equation? The answer is given in the following theorem: Theorem (Test for Exactness) Let M, N, and their first partial derivatives My and Nx , be continuous in a simply connected domain D of the xy -plane. Then the differential equation M(x, y )dx + N(x, y )dy = 0 is exact for all (x, y ) in D if and only if ∂M ∂N = , that is My = Nx . ∂y ∂x MATH2030 Differential Equations (AY 2023-24) 132 / 177 Here a simply connected domain D is a domain such that every simple closed contour within it encloses only points of D. The set of points interior to a simple closed contour is an example. The annular domain between two concentric circles is not simply connected. A domain that is not simply connected is said to be multiply connected. Remark Since there is no need to memorize the final result for u, we omit the details of the proof of the theorem above. For a particular problem one can construct an appropriate potential functions from first principles. MATH2030 Differential Equations (AY 2023-24) 133 / 177 Examples Determine whether the given differential equation is exact. x y dy y 3 ) dy 1 [1 + log(xy )] dx + = 0. 2 x 2 y dx − (xy 2 + = 0. Solution 1 M(x, y ) = [1 + log(xy )] =⇒ M = y We have N(x, y ) = x =⇒ N = 1 . y 2 x 1 y y So My = Nx and the differential equation is exact. M(x, y ) = x 2 y =⇒ M = x 2 y We have N(x, y ) = −(xy 2 + y 3 ) =⇒ N = −y 2 . x Thus My 6= Nx , and hence the differential equation is not exact. MATH2030 Differential Equations (AY 2023-24) 134 / 177