π Lecture 9 β Differential Equations (MTH401) β Bernoulli Equations and Substitutions
π Overview: This lecture introduces Bernoulli differential equations, a nonlinear form that can be transformed into a linear differential equation through a specific substitution. It also covers the use of various creative substitutions to solve differential equations that do not fit standard forms, including several solved examples and exercises.
ποΈ Topics Covered
The lecture begins with the definition and method of solution for Bernoulli equations, including the substitution v = y^(1-n) to convert them to linear form. It then provides worked examples of solving Bernoulli equations. Following this, the lecture covers the technique of using arbitrary substitutions to transform non-standard differential equations into solvable forms like separable, homogeneous, exact, or linear equations, illustrated with several examples. The section concludes with exercises and a set of solved problems covering homogeneous, exact, and linear equations.
π Lecture Summary
8 Bernoulli Equations
A differential equation that can be written in the form dy/dx + p(x)y = q(x)y^n is called a Bernoulli equation.
8.1 Method of solution
For n = 0 or n = 1, the equation reduces to a 1st order linear DE and can be solved accordingly. For n β 0, 1, we divide the equation with y^n to write it in the form y^(-n) dy/dx + p(x) y^(1-n) = q(x) and then put v = y^(1-n). Differentiating w.r.t. βxβ, we obtain vβ = (1-n) y^(-n) yβ. Therefore, the equation becomes dv/dx + (1-n) p(x) v = (1-n) q(x). This is a linear equation satisfied by v. Once it is solved, you will obtain the function y = v^(1/(1-n)). If n > 1, then we add the solution y = 0 to the solutions found by the above technique.
Summary
- Identify the equation as Bernoulli equation. Find n. If n β 0, 1, divide by y^n and substitute v = y^(1-n).
- Through easy differentiation, find the new equation dv/dx + (1-n) p(x) v = (1-n) q(x).
- This is a linear equation. Solve the linear equation to find v.
- Go back to the old function y through the substitution y = v^(1/(1-n)).
- If n > 1, then include y = 0 in the solution.
- If you have an IVP, use the initial condition to find the particular solution.
Example 1: Solve the equation dy/dx = y + y^3
- The given differential can be written as dy/dx - y = y^3, which is a Bernoulli equation with p(x) = -1, q(x) = 1, n = 3. Dividing with y^3, we get y^(-3) dy/dx - y^(-2) = 1. Therefore we substitute v = y^(1-3) = y^(-2).
- Differentiating w.r.t. βxβ we have y^(-3) dy/dx = -(1/2)(dv/dx). So that the equation reduces to dv/dx + 2v = -2.
- This is a linear equation. To solve this we find the integrating factor u(x) = e^(β«2dx) = e^(2x). The solution of the linear equation is given by v = [β« u(x)q(x) dx + c] / u(x). Since β« e^(2x)(-2)dx = -e^(2x), the solution for v is given by v = (-e^(2x) + c) / e^(2x) = ce^(-2x) - 1.
- To go back to y we substitute v = y^(-2). Therefore the general solution of the given DE is y^(-2) = ce^(-2x) - 1 or y = Β± (ce^(-2x) - 1)^(-1/2).
- Since n > 1, we include y = 0 in the solutions. Hence, all solutions are y = 0, y = Β± (ce^(-2x) - 1)^(-1/2).
π Definition β Bernoulli Equation: A first-order differential equation of the form dy/dx + p(x)y = q(x)y^n, where n is any real number. π Formula: v = y^(1-n) β Substitution to convert a Bernoulli equation into a linear equation in v. π Example: dy/dx - y = y^3. Here n=3, so v = y^(1-3) = y^(-2). The resulting linear equation is dv/dx + 2v = -2, with integrating factor u(x)=e^(2x). The final general solution is y^(-2) = ce^(-2x) - 1.
Example 2: Solve dy/dx + (1/x) y = xy^2 In the given equation we identify p(x) = 1/x, q(x) = x, and n = 2. Thus the substitution w = y^(-1) gives dw/dx - (1/x) w = -x. The integrating factor for this linear equation is u(x) = e^(β« -1/x dx) = x^(-1). Multiplying gives x^(-1) w = β« -x(x^(-1)) dx = β« -1 dx = -x + c, or w = -x^2 + cx. Since w = y^(-1), we obtain y = 1 / (cx - x^2). For n > 0 the trivial solution y = 0 is a singular solution of the given equation.
Example 3: Solve dy/dx + (x/(1-x^2)) y = x y^(1/2) Dividing by y^(1/2), the given equation becomes y^(-1/2) dy/dx + (x/(1-x^2)) y^(1/2) = x. Put y^(1/2) = v. Then (1/2) y^(-1/2) dy/dx = dv/dx, so y^(-1/2) dy/dx = 2 dv/dx. The equation reduces to 2 dv/dx + (x/(1-x^2)) v = x, or dv/dx + (x/(2(1-x^2))) v = x/2. This is linear in v. The integrating factor is e^(β« x/(2(1-x^2)) dx) = (1-x^2)^(-1/4). Multiplying by the I.F. gives d/dx [v(1-x^2)^(-1/4)] = (x/2)(1-x^2)^(-1/4). Integrating, we get v(1-x^2)^(-1/4) = -(1/3)(1-x^2)^(3/4) + c. So v = c(1-x^2)^(1/4) - (1/3)(1-x^2). Since v = y^(1/2), the required solution is y^(1/2) = c(1-x^2)^(1/4) - (1/3)(1-x^2).
8.2 Exercise
Solve the following differential equations:
- x dy/dx + y = y^2 ln x
- dy/dx + y = xy^3
- dy/dx - y = e^x y^2
- dy/dx = y (xy^3 - 1)
- x dy/dx - (1+x)y = xy^2
- x^2 dy/dx + y^2 = xy Solve the initial-value problems:
- x^2 dy/dx - 2xy = 3y^4, y(1) = 1/2
- dy/dx = y/x - y^2/x^2, y(-1) = 1
- 2y^(1/2) dy/dx + y^(3/2) = 1, y(0) = 4
- y(1+2xy) dx + x(1-2xy) dy = 0
8.3 Substitutions
Sometimes a differential equation can be transformed by means of a substitution into a form that could then be solved by one of the standard methods i.e. Methods used to solve separable, homogeneous, exact, linear, and Bernoulliβs differential equation. An equation may look different from any of those that we have studied in the previous lectures, but through a sensible change of variables perhaps an apparently difficult problem may be readily solved. Although no firm rules can be given on the basis of which these substitution could be selected, a working axiom might be: Try something! It sometimes pays to be clever.
Example 1 The differential equation 2x^2 dy/dx + y^2 = xy is not separable, not homogeneous, not exact, not linear, and not Bernoulli. However, we might try the substitution u = 2xy or y = u/(2x). The equation becomes, after we simplify, 2u^2 dx + (1-u)x du = 0. Separating and integrating gives 2 ln x - u^(-1) - ln u = c. Re-substituting u = 2xy, we obtain ln x = c + 1/(2y) - ln(2xy), which simplifies to x = 2c1 y e^(1/(2xy)), where e^c was replaced by c1. The differential equation also possesses the trivial solution y=0, but this function is not included in the one-parameter family of solution.
π‘ Why this matters: When faced with a non-standard differential equation, a creative substitution can transform it into a form we know how to solve, such as a separable equation.
Example 2: Solve x dy/dx + 2y = 3x - 6 The presence of the term 2y dy/dx prompts us to try u = y^2. Since du/dx = 2y dy/dx, the equation becomes x(du/dx)/2 + 2u = 3x - 6, or x du/dx + 4u = 6x - 12, then dividing by x: du/dx + (4/x) u = 6 - 12/x. This equation has the form of a 1st order linear differential equation with P(x) = 4/x and Q(x) = 6 - 12/x. The integrating factor is I.F. = e^(β« 4/x dx) = x^4. Multiplying gives d/dx [x^4 u] = 6x^4 - 12x^3. Integrating both sides, we obtain x^4 u = (6/5)x^5 - 3x^4 + c, or u = (6/5)x - 3 + c/x^4. Re-substituting u = y^2, the general solution is y^2 = (6/5)x - 3 + c/x^4.
Example 3: Solve x - y = e^(y/x) x^3 dy/dx If we let u = y/x or y = ux, then dy/dx = u + x du/dx. The given differential equation becomes x - ux = e^u x^3 (u + x du/dx), which simplifies to ue^(-u) du = dx. Integrating both sides, we have β« u e^(-u) du = β« dx. Using integration by parts, we get -ue^(-u) - e^(-u) = x + c, or u + 1 = (c1 - x) e^u, where c1 = -c. We then re-substitute u = y/x and simplify to obtain y + x = x(c1 - x) e^(y/x).
Example 4: Solve dy/dx = 2x (dy/dx)^2 If we let u = dy/dx, then du/dx = d^2y/dx^2. The equation reduces to du/dx = 2x u^2, which is separable. Separating the variables, we obtain du / u^2 = 2x dx. Integrating both sides yields -1/u = x^2 + c2, or u = -1/(x^2 + c2). Since u = dy/dx, we integrate again: β« dy = -β« dx/(x^2 + c2). This gives y + c = -(1/βc2) tan^(-1)(x/βc2).
8.4 Exercise
Solve the differential equations by using an appropriate substitution.
- y dx + (1 + y e^x) dy = 0
- (2 + e^(-x/y)) dx + 2(1 - x/y) dy = 0
- 2x csc(2y) dy/dx = 2x - ln(tan y)
- dy/dx + 1 = sin x e^(-(x+y))
- y dy/dx + 2x ln x = x e^y
- x^2 dy/dx + 2xy = x^4 y^2 + 1
- x e^y dy/dx - 2e^y dx = x^2
9 Solved Problems
Example 1: yβ = (x^2 + y^2) / (xy) This is a homogeneous equation. Put y = wx, then dy/dx = w + x dw/dx. Substituting, we get w + x dw/dx = (x^2 + w^2 x^2) / (xwx) = (1 + w^2)/w. This simplifies to x dw/dx = 1/w, or w dw = dx/x. Integrating gives w^2/2 = ln|x| + c, or (y/x)^2/2 = ln|x| + c, so y^2 = 2x^2 ln|cx|.
Example 2: dy/dx = (2xy - y) / x This is a homogeneous equation. Put y = wx, then dy/dx = w + x dw/dx. Substituting, we get w + x dw/dx = (2x(wx) - wx)/x = 2w - w = w. This simplifies to x dw/dx = 0, so dw/dx = 0. Integrating gives w = c, so y = cx. Let's re-check: w + x dw/dx = (2x(wx) - wx)/x = (2wx^2 - wx)/x = 2wx - w = w(2x - 1). This is incorrect in the text. The correct simplification: (2β(xy) - y)/x. Let u = β(xy). Actually, the text's solution is: w + x dw/dx = (2β(x(wx)) - wx)/x = (2xβw - wx)/x = 2βw - w. So x dw/dx = 2βw - 2w. Separating: dw/(2βw - 2w) = dx/x. This simplifies to dw/(2βw(1 - βw)) = dx/x. Let t = βw, then dt = dw/(2βw), so dt/(1-t) = dx/x. Integrating: -ln|1-t| = ln|x| + ln|c|, so 1/(1-t) = cx. Therefore t = 1 - 1/(cx), and since t = βw = β(y/x), we get β(y/x) = 1 - 1/(cx).
Example 3: (2y^2 x - 3) dx + (2yx^2 + 4) dy = 0 Check for exactness: M = 2xy^2 - 3, N = 2x^2 y + 4. βM/βy = 4xy, βN/βx = 4xy. The equation is exact. Find f(x,y): β« M dx = β« (2xy^2 - 3) dx = x^2 y^2 - 3x + h(y). Then βf/βy = 2x^2 y + h'(y) = 2x^2 y + 4 = N, so h'(y) = 4, and h(y) = 4y + c. The solution is x^2 y^2 - 3x + 4y = C.
Example 4: dy/dx = (2xy e^(x/y)^2) / (y^2 + y^2 e^(x/y)^2 + 2x^2 e^(x/y)^2) This requires the substitution x/y = w, which simplifies to a separable equation. After substitution and integration, the solution is y = c(1 + e^(x/y)^2).
Example 5: x dy/dx + y/(ln x) = 3x^2/(ln x) Rewrite as dy/dx + (1/(x ln x)) y = 3x/ln x. This is a linear equation. The integrating factor is I.F. = e^(β« 1/(x ln x) dx) = e^ln|ln x| = ln x. Multiplying, we get d/dx (y ln x) = 3x. Integrating gives y ln x = 3x^2/2 + c, so y = (3x^2/2 + c) / ln x.
Example 6: (y^2 e^x + 2xy) dx - x^2 dy = 0 Check for exactness: M = y^2 e^x + 2xy, N = -x^2. βM/βy = 2y e^x + 2x, βN/βx = -2x. The equation is not exact. Divide by y^2: (e^x + 2x/y) dx - (x^2/y^2) dy = 0. Now βM/βy = -2x/y^2, βN/βx = -2x/y^2. The equation is exact. Integrate: f(x,y) = β« (e^x + 2x/y) dx = e^x + x^2/y + h(y). βf/βy = -x^2/y^2 + h'(y) = -x^2/y^2, so h'(y) = 0, h(y)=c. The solution is e^x + x^2/y = C.
β Key Takeaways
A student must master the identification of Bernoulli equations (dy/dx + p(x)y = q(x)y^n) and the correct substitution method (v = y^(1-n)) to convert them into solvable linear equations, remembering to include the trivial solution y=0 when n > 1. The power of creative substitutions is a critical tool for tackling non-standard differential equations that don't fit the standard forms, but no general rule exists beyond trial and error. Recognizing exact equations and making them exact by multiplying by an appropriate factor is a valuable technique. Finally, practicing the complete solution processβfrom identification and substitution to solving the transformed equation and back-substitutingβis essential, including using initial conditions for particular solutions in Initial Value Problems.
π§ Quick Revision Questions
- What is the general form of a Bernoulli differential equation, and what substitution converts it into a linear differential equation?
- A differential equation is reduced to dv/dx + (2/x) v = x. What is its integrating factor, and what is the general solution for v?
- In Example 1 of Section 8.3, why is the substitution u = 2xy suggested, and what type of equation does it transform the original equation into?
- For the initial value problem x^2 dy/dx - 2xy = 3y^4 with y(1) = 1/2, what is the first step and the integrating factor for the resulting linear equation in v?
- Describe the key difference in the solution process for a Bernoulli equation when n > 1 compared to when n = 0 or n = 1.
π Lecture 10 β Differential Equations (MTH401)
π Overview: This lecture focuses on solving first-order differential equations using substitution methods to reduce them to standard forms like linear, Bernoulli, or separable equations. It demonstrates how strategic substitutions can simplify complex differential equations into solvable forms, a critical skill for tackling non-standard ODEs.
ποΈ Topics Covered
The lecture presents a series of solved examples (Examples 7 through 17) illustrating various substitution techniques. These include reducing equations to linear form using integrating factors, applying substitutions like ( e^{2y} = u ), (\ln y = u), (\ln \tan y = u), and (x+y=u) to simplify and solve differential equations, along with examples involving Bernoulli's equation and specific function substitutions like (x^3y^3 = u).
π Lecture Summary
Example 7: Solve ( x \cos x \frac{dy}{dx} + y(x \sin x + \cos x) = 1 )
The equation is first divided by (x \cos x) to put it in standard linear form: ( \frac{dy}{dx} + y \left[ \frac{x \sin x + \cos x}{x \cos x} \right] = \frac{1}{x \cos x} ). This simplifies to ( \frac{dy}{dx} + y \left[ \tan x + \frac{1}{x} \right] = \frac{1}{x \cos x} ). The integrating factor (I.F.) is calculated as ( \text{I.F.} = \exp\left(\int (\tan x + \frac{1}{x}) dx\right) = x \sec x ). Multiplying through gives ( \frac{d}{dx} [xy \sec x] = \sec^2 x ). Integrating yields the solution: ( xy \sec x = \tan x + c ).
π Definition β Integrating Factor (I.F.): A function used to multiply a non-exact differential equation to make it exact and integrable. π Formula: ( \mu(x) = e^{\int P(x) dx} ) for a linear ODE ( \frac{dy}{dx} + P(x)y = Q(x) ), which simplifies to ( \frac{d}{dx}[\mu y] = \mu Q(x) ). π Example: For the simplified ODE, (P(x) = \tan x + 1/x) and (Q(x) = 1/(x \cos x)). The I.F. is (e^{\int (\tan x + 1/x) dx} = e^{\ln|\sec x| + \ln|x|} = x \sec x). The left side becomes the derivative of (xy\sec x), leading to the final solution.
Example 8: Solve ( x e^{2y} \frac{dy}{dx} + e^{2y} = \frac{\ln x}{x} )
Use the substitution ( e^{2y} = u ). Differentiating gives ( 2e^{2y} \frac{dy}{dx} = \frac{du}{dx} ), so ( e^{2y} \frac{dy}{dx} = \frac{1}{2} \frac{du}{dx} ). Substituting transforms the equation into ( x \left(\frac{1}{2} \frac{du}{dx}\right) + u = \frac{\ln x}{x} ), or ( \frac{du}{dx} + \frac{2}{x} u = \frac{2 \ln x}{x^2} ). This is linear in (u). The I.F. is ( e^{\int (2/x) dx} = x^2 ). Multiplying gives ( \frac{d}{dx}(x^2 u) = 2 \ln x ). Integrating results in ( x^2 u = 2[x \ln x - x] + c ). Substituting back ( u = e^{2y} ) gives the implicit solution: ( x^2 e^{2y} = 2[x \ln x - x] + c ).
Example 9: Solve ( \frac{dy}{dx} + y \ln y = y e^x )
Divide both sides by (y) (assuming (y \neq 0)) to get ( \frac{1}{y} \frac{dy}{dx} + \ln y = e^x ). Use the substitution ( \ln y = u ). Then ( \frac{1}{y} \frac{dy}{dx} = \frac{du}{dx} ). The equation becomes ( \frac{du}{dx} + u = e^x ). This is a linear equation in (u). The I.F. is ( e^{\int 1 dx} = e^x ). Multiplying yields ( \frac{d}{dx}(e^x u) = e^{2x} ). Integrating gives ( e^x u = \frac{e^{2x}}{2} + c ). Substituting back ( u = \ln y ) yields the solution: ( e^x \ln y = \frac{e^{2x}}{2} + c ).
Example 10: Solve ( 2x \csc 2y \frac{dy}{dx} = 2x - \ln \tan y )
Use the substitution ( \ln \tan y = u ). Differentiating gives ( \frac{dy}{dx} = \sin y \cos y \frac{du}{dx} ). The original equation becomes ( 2x \csc 2y (\sin y \cos y \frac{du}{dx}) = 2x - u ). Since ( \csc 2y = 1/\sin 2y ) and ( \sin 2y = 2 \sin y \cos y ), this simplifies to ( x \frac{du}{dx} = 2x - u ). Rearranging gives ( \frac{du}{dx} + \frac{1}{x} u = 2 ). The I.F. is ( e^{\int (1/x) dx} = x ). Multiplying gives ( \frac{d}{dx}(xu) = 2x ). Integrating yields ( xu = x^2 + c ), so ( u = x + c x^{-1} ). Substituting back: ( \ln \tan y = x + c x^{-1} ).
Example 11: Solve ( \frac{dy}{dx} + x + y + 1 = (x + y)^2 e^{3x} )
Use the substitution ( x+y = u ). Then ( dy/dx = du/dx - 1 ). The equation becomes ( (du/dx - 1) + u + 1 = u^2 e^{3x} ), which simplifies to ( \frac{du}{dx} + u = u^2 e^{3x} ). This is Bernoulli's equation. Rewrite as ( u^{-2} \frac{du}{dx} + u^{-1} = e^{3x} ). Let ( w = u^{-1} ). Then ( dw/dx = -u^{-2} du/dx ), so ( -dw/dx + w = e^{3x} ) or ( \frac{dw}{dx} - w = -e^{3x} ). This is linear in (w). The I.F. is ( e^{\int -1 dx} = e^{-x} ). Multiplying: ( d/dx (e^{-x} w) = -e^{2x} ). Integrating: ( e^{-x} w = -e^{2x}/2 + c ), so ( w = -e^{3x}/2 + c e^{x} ). Since ( w = 1/u ), ( 1/u = -e^{3x}/2 + c e^{x} ). Substituting ( u = x+y ) gives ( \frac{1}{x+y} = -\frac{e^{3x}}{2} + c e^{x} ).
π Definition β Bernoulli's Equation: A differential equation of the form ( \frac{dy}{dx} + P(x)y = Q(x) y^n ). π Formula: Solve by dividing by (y^n) and substituting ( w = y^{1-n} ), which transforms it into a linear equation in (w). π Example: Here, ( \frac{du}{dx} + u = u^2 e^{3x} ) is Bernoulli with (n=2). Dividing by (u^2) and substituting (w = u^{-1}) yields the linear ODE ( \frac{dw}{dx} - w = -e^{3x} ).
Example 12: Solve ( \frac{dy}{dx} = (4x + y + 1)^2 )
Use the substitution ( 4x + y + 1 = u ). Then ( dy/dx = du/dx - 4 ). The equation becomes ( du/dx - 4 = u^2 ), or ( du/dx = u^2 + 4 ). This is separable. Separate: ( du/(u^2 + 4) = dx ). Integrate: ( \frac{1}{2} \tan^{-1}(u/2) = x + c ). Multiplying by 2: ( \tan^{-1}(u/2) = 2x + c_1 ). So ( u = 2 \tan(2x + c_1) ). Substituting back: ( 4x + y + 1 = 2 \tan(2x + c_1) ).
Example 13: Solve ( (x + y)^2 \frac{dy}{dx} = a^2 )
Use the substitution ( x+y = u ). Then ( dy/dx = du/dx - 1 ). The equation becomes ( u^2 (du/dx - 1) = a^2 ), or ( u^2 du/dx = a^2 + u^2 ). Rearranging: ( du = (a^2 + u^2)/u^2 dx ), or ( dx = u^2/(a^2 + u^2) du ). Simplify the integrand: ( u^2/(a^2 + u^2) = 1 - a^2/(a^2 + u^2) ). Integrating: ( x + c = \int (1 - a^2/(a^2 + u^2)) du = u - a \tan^{-1}(u/a) ). Substituting back ( u = x+y ): ( x + c = (x+y) - a \tan^{-1}\left(\frac{x+y}{a}\right) ). Rearranged: ( y - a \tan^{-1}\left(\frac{x+y}{a}\right) = c ). (Note: in the text, the final answer is given as ( (x+y) - a \tan^{-1}((x+y)/a) = x + c ), which simplifies to this form.)
Example 14: Solve ( 2y \frac{dy}{dx} + x^2 + y^2 + x = 0 )
Use the substitution ( x^2 + y^2 = u ). Differentiating: ( 2x + 2y \frac{dy}{dx} = \frac{du}{dx} ), or ( 2y \frac{dy}{dx} = \frac{du}{dx} - 2x ). Substituting into the equation: ( (\frac{du}{dx} - 2x) + x^2 + y^2 + x = 0 ). But ( x^2 + y^2 = u ), so ( \frac{du}{dx} - 2x + u + x = 0 ), which simplifies to ( \frac{du}{dx} + u = x ). This is linear in (u). The I.F. is ( e^{\int 1 dx} = e^x ). Multiplying: ( d/dx(e^x u) = x e^x ). Integrating by parts: ( e^x u = x e^x - e^x + c ). Substituting back ( u = x^2 + y^2 ): ( e^x (x^2 + y^2) = x e^x - e^x + c ), or ( x^2 + y^2 = x - 1 + c e^{-x} ).
Example 15: Solve ( y' + 1 = e^{-(x+y)} \sin x )
Use the substitution ( x + y = u ). Then ( y' = du/dx - 1 ). Substituting: ( (du/dx - 1) + 1 = e^{-u} \sin x ), which simplifies to ( du/dx = e^{-u} \sin x ). This is separable. Rearranging: ( e^u du = \sin x dx ). Integrate: ( e^u = -\cos x + c ). Taking the natural log: ( u = \ln |c - \cos x| ). Substituting back ( u = x + y ): ( x + y = \ln |c - \cos x| ).
Example 16: Solve ( x^4 y^2 y' + x^3 y^3 = 2x^3 - 3 )
Use the substitution ( x^3 y^3 = u ). Differentiating: ( 3x^2 y^3 + 3x^3 y^2 y' = \frac{du}{dx} ). Rearranging: ( 3x^3 y^2 y' = \frac{du}{dx} - 3x^2 y^3 ), so ( x^4 y^2 y' = \frac{x}{3} \frac{du}{dx} - x^3 y^3 ). But ( x^3 y^3 = u ), so the left side of the original equation becomes ( \frac{x}{3} \frac{du}{dx} - u ). Substituting into the original ODE: ( \frac{x}{3} \frac{du}{dx} - u + u = 2x^3 - 3 ), which simplifies to ( \frac{x}{3} \frac{du}{dx} = 2x^3 - 3 ). Rearranging: ( \frac{du}{dx} = 6x^2 - \frac{9}{x} ). Integrating: ( u = 2x^3 - 9 \ln x + c ). Substituting back ( u = x^3 y^3 ): ( x^3 y^3 = 2x^3 - 9 \ln x + c ).
Example 17: Solve ( \cos(x + y) dy = dx )
Use the substitution ( x + y = v ). Then ( 1 + dy/dx = dv/dx ), or ( dy/dx = dv/dx - 1 ). The original equation can be written as ( \cos(v) dy = dx ), or ( dy = dx / \cos(v) ). Since ( dy = dv - dx ), we get ( dv - dx = dx / \cos(v) ). Rearranging: ( dv = dx (1 + 1/\cos(v)) = dx ((\cos(v) + 1)/\cos(v)) ). Thus, ( dx = \frac{\cos(v)}{1 + \cos(v)} dv ). Simplify the integrand: ( \frac{\cos v}{1 + \cos v} = 1 - \frac{1}{1 + \cos v} ). Using the half-angle identity, ( 1 + \cos v = 2 \cos^2(v/2) ), so ( 1/(1+\cos v) = \frac{1}{2} \sec^2(v/2) ). Thus, ( dx = [1 - \frac{1}{2} \sec^2(v/2)] dv ). Integrating: ( x + c = v - \tan(v/2) ). Substituting back ( v = x+y ): ( x + c = (x + y) - \tan\left(\frac{x+y}{2}\right) ). Simplifying: ( y = \tan\left(\frac{x+y}{2}\right) + c ). π‘ Why this matters: This example shows how trigonometric substitutions and identities can be combined to solve an implicit ODE.
β Key Takeaways
The most critical skills from this lecture are recognizing when a first-order ODE can be simplified by a clever substitution, such as (x+y), (e^{2y}), (\ln y), or (x^2 + y^2), to transform it into a linear, Bernoulli, or separable equation. Mastery involves identifying the type of substitution based on the structure of the equation (e.g., terms like (x+y) or (\ln y) or (e^{2y}) hint at their use) and then carefully applying the chain rule to rewrite the derivative. For exam success, students must be fluent in computing integrating factors for linear ODEs and solving Bernoulli equations with the standard (y^{1-n}) substitution, and they must practice these techniques to avoid algebraic errors.
π§ Quick Revision Questions
- What substitution would you use to solve (\frac{dy}{dx} + y \ln y = y e^x) and why?
- How does the substitution (x+y=u) transform the term (\frac{dy}{dx}) in an ODE?
- What is the standard method for solving a Bernoulli equation of the form (\frac{dy}{dx} + P(x)y = Q(x) y^n)?
- For Example 8, what is the exact substitution used and how does it help eliminate terms involving (e^{2y})?
- In Example 12, after substituting (4x+y+1=u), the equation becomes separable. Write down the separated form (f(u) du = dx).
π Lecture 10 β Applications of First Order Differential Equations
π Overview: This lecture introduces how differential equations model physical phenomena, focusing on finding differential equations from families of curves and the concept of orthogonal trajectories. Understanding orthogonal trajectories is crucial for applications in physics, electromagnetism, and fluid dynamics where curves intersect at right angles.
ποΈ Topics Covered
The lecture covers the basic steps of mathematical modeling for physical phenomena, finding differential equations from families of curves by eliminating parameters, defining orthogonal curves and orthogonal trajectories, and presenting a systematic method for finding orthogonal trajectories of a given family of curves through differential equations.
π Lecture Summary
Step 1: State Assumptions
We clearly state the assumptions on which the model will be based. These assumptions should describe the relationships among the quantities to be studied.
Step 2: Identify Parameters and Variables
Completely describe the parameters and variables to be used in the model.
Step 3: Derive Equations
Use the assumptions (from Step 1) to derive mathematical equations relating the parameters and variables (from Step 2).
The mathematical models for physical phenomenon often lead to a differential equation or a set of differential equations. The applications of the differential equations discussed in this lecture include:
- Orthogonal Trajectories
- Population dynamics
- Radioactive decay
- Newtonβs Law of cooling
- Carbon dating
- Chemical reactions
Finding Differential Equations from Families of Curves
We know that solutions of a 1st order differential equation, e.g. separable equations, may be given by an implicit equation F(x, y, C) = 0 with 1 parameter C, representing a family of curves. Similarly, an nth order DE yields an n-parameter family of curves.
The question arises whether we can turn the problem around: Starting with an n-parameter family of curves, can we find an associated nth order differential equation free of parameters? The answer in most cases is yes.
Procedure for finding DE from a 1-parameter family:
- Differentiate with respect to x to get an equation involving x, y, dy/dx, and C.
- Use the original equation to eliminate the parameter C from the new equation.
- Use algebra to rewrite this equation in explicit form dy/dx = f(x, y).
Example: Find the differential equation satisfied by the family xΒ² + yΒ² = Cx
- Differentiate with respect to x: 2x + 2y dy/dx = C
- From the original equation: C = (xΒ² + yΒ²)/x
- Substitute: 2x + 2y dy/dx = (xΒ² + yΒ²)/x
This gives: dy/dx = (yΒ² - xΒ²)/(2xy)
10.1 Orthogonal Trajectories
Two families of curves example:
- First family: y = mx (all straight lines through origin)
- Second family: xΒ² + yΒ² = C (all circles centered at origin)
When these two families are drawn together, whenever one line intersects one circle, the tangent line to the circle and the line are perpendicular, i.e., orthogonal to each other. The two families of curves are orthogonal at the point of intersection.
10.2 Orthogonal Curves
Any two curves Cβ and Cβ are said to be orthogonal if their tangent lines Tβ and Tβ at their point of intersection are perpendicular. This means that slopes are negative reciprocals of each other, except when Tβ and Tβ are parallel to the coordinate axes.
10.3 Orthogonal Trajectories (OT)
When all curves of a family Ξβ: G(x, y, cβ) = 0 orthogonally intersect all curves of another family Ξβ: H(x, y, cβ) = 0, then each curve of the families is said to be an orthogonal trajectory of the other.
Orthogonal trajectories occur naturally in many areas of physics, fluid dynamics, and the study of electricity and magnetism. For example, the lines of force are perpendicular to the equipotential curves (curves of constant potential).
10.3.1 Method of Finding Orthogonal Trajectories
Consider a family of curves Ξ. Assume an associated DE is found: dy/dx = f(x, y)
Since dy/dx gives the slope of the tangent to a curve of the family Ξ through (x,y), the slope of the line that is tangent to the orthogonal curve through (x,y) is given by -1/f(x,y). The family of orthogonal curves are solutions to the differential equation: dy/dx = -1/f(x,y)
Summary β Steps for finding Orthogonal Trajectories:
- Step 1: Consider a family of curves Ξ and find the associated differential equation.
- Step 2: Rewrite this differential equation in the explicit form dy/dx = f(x,y).
- Step 3: Write down the differential equation associated to the orthogonal family: dy/dx = -1/f(x,y).
- Step 4: Solve the new equation. The solutions are exactly the family of orthogonal curves.
- Step 5: If a specific curve from the orthogonal family is required, solve an IVP.
Example 1
Find the Orthogonal Trajectory to the family of circles xΒ² + yΒ² = CΒ²
Solution: The given equation represents a family of concentric circles centered at the origin.
Step 1: Differentiate w.r.t. x to find the DE: 2x + 2y dy/dx = 0 βΉ dy/dx = -x/y
Step 2: Rewrite in explicit form: dy/dx = -x/y
Step 3: DE for orthogonal family: dy/dx = y/x
Step 4: This is a separable DE. Using integrating factor for linear equation: u(x) = e^β«(-1/x)dx = 1/x
Solution: y Β· u(x) = m βΉ y Β· (1/x) = m βΉ y = mx
This represents a family of straight lines through the origin. Hence the family of straight lines y = mx and the family of circles xΒ² + yΒ² = CΒ² are Orthogonal Trajectories.
Example 2
Find the Orthogonal Trajectory to the family of circles xΒ² + yΒ² = 2Cx
Solution:
Step 1: Differentiate: 2x + 2y dy/dx = 2C From original: C = (xΒ² + yΒ²)/(2x)
Substitute: 2x + 2y dy/dx = 2[(xΒ² + yΒ²)/(2x)] = (xΒ² + yΒ²)/x
Step 2: Explicit DE: dy/dx = (yΒ² - xΒ²)/(2xy)
Step 3: DE for orthogonal family: dy/dx = -2xy/(yΒ² - xΒ²)
[The solution continues but is not completed in the provided text.]
β Key Takeaways
The key concept is that orthogonal trajectories are families of curves that intersect at right angles, and their slopes are negative reciprocals at intersection points. To find orthogonal trajectories, first determine the differential equation of the given family, then replace dy/dx with -1/(dy/dx) and solve the resulting equation. This method applies to concentric circles giving straight lines through the origin, and more complex families require similar parameter elimination steps. The process involves differentiation, parameter elimination, forming the orthogonal DE, and solving it.
π§ Quick Revision Questions
- What is the condition for two curves to be orthogonal at their point of intersection?
- What are the four steps to find orthogonal trajectories of a family of curves?
- What is the orthogonal trajectory family for concentric circles centered at the origin?
- How do you eliminate the parameter C when finding a DE from a family of curves?
- Give two physical applications where orthogonal trajectories occur naturally.
π Lecture 12 β Exact Lecture Title: Differential Equations (MTH401) β Lecture 12
π Overview: This lecture continues the detailed solution of a homogeneous differential equation from the previous lecture, then introduces real-world applications of first-order differential equations. It focuses on population dynamics (exponential growth), radioactive decay, and Newton's Law of Cooling, models that are foundational in biology, physics, and engineering for predicting change over time.
ποΈ Topics Covered
The lecture begins by completing the solution of a homogeneous differential equation from the previous lecture, using a substitution ( y = vx ) to reduce it to a separable form. The remainder of the lecture is dedicated to applying first-order differential equations to three key areas: population dynamics with exponential growth, radioactive decay using the concept of half-life, and Newton's Law of Cooling.
π Lecture Summary
4. [Continuation from previous page: Solving the Homogeneous DE]
This DE is a homogeneous equation. To solve it, we substitute ( y = vx ), or equivalently ( v = y/x ). Then we have ( \frac{dy}{dx} = x \frac{dv}{dx} + v ). Therefore the homogeneous differential equation from step 3 becomes ( x \frac{dv}{dx} + v = \frac{2v}{1 - v^2} ). Algebraic manipulations reduce this equation to the separable form: ( \frac{v + v^3}{1 - v^2} dv = \frac{1}{x} dx ).
The constant solutions are given by ( v + v^3 = 0 \Rightarrow v(1 + v^2) = 0 ). The only constant solution is ( v = 0 ). To find the non-constant solutions we separate the variables: ( \int \frac{v + v^3}{1 - v^2} dv = \int \frac{1}{x} dx ). Resolving into partial fractions the integrand on LHS, we obtain: ( \frac{v + v^3}{1 - v^2} = \frac{1}{v} - \frac{2v}{1 + v^2} ). Hence we have: ( \int \left( \frac{1}{v} - \frac{2v}{1 + v^2} \right) dv = \int \frac{1}{x} dx ). Hence the solution of the separable equation becomes: ( \ln |v| - \ln [v^2 + 1] = \ln |x| + \ln C ), where ( C \neq 0 ). This is equivalent to ( \frac{v}{v^2 + 1} = C x ). We go back to ( y ) to get: ( \frac{y/x}{(y/x)^2 + 1} = C x \Rightarrow \frac{y}{x} = C x \left( \frac{y^2}{x^2} + 1 \right) ), and which is equivalent to: The final set of solutions is: ( v = 0 \Rightarrow y = 0 ) (x-axis) and ( \frac{v}{v^2 + 1} = C x \Rightarrow \frac{y/x}{(y/x)^2 + 1} = Cx \Rightarrow x^2 + y^2 = m y ), where ( m = 1/C ). This is a family of circles centered on the y-axis.
10.4 Population Dynamics
Some natural questions related to population problems are the following: What will the population of a certain country be after e.g. ten years? How are we protecting the resources from extinction? The easiest population dynamics model is the exponential model. This model is based on the assumption: The rate of change of the population is proportional to the existing population.
If ( P(t) ) measures the population of a species at any time ( t ), then because of the above-mentioned assumption we can write ( \frac{dP}{dt} = kP ). The rate ( k ) is a constant of proportionality. Clearly, the above equation is linear as well as separable. To solve this equation, we multiply the equation with the integrating factor ( e^{-kt} ) to obtain ( \frac{d}{dt} [P e^{-kt}] = 0 ). Integrating both sides we obtain ( P e^{-kt} = C ) or ( P = C e^{kt} ). Clearly, we must have ( k > 0 ) for growth and ( k < 0 ) for the decay. If ( P_0 ) is the initial population then ( P(0) = P_0 ). So that ( C = P_0 ) and we obtain ( P(t) = P_0 e^{kt} ). π‘ Why this matters: This model, while simple, forms the basis for understanding unrestrained growth (like bacteria in a petri dish) and decay (like drug concentration in the body).
π Definition β Exponential Model: A population model where the rate of change of population is directly proportional to the current population. π Formula: ( P(t) = P_0 e^{kt} ) β The population at time ( t ) equals the initial population multiplied by Euler's number raised to the product of the growth rate and time. π Example: A community's population has doubled in 5 years. Find the growth rate constant ( k ). Solution: ( 2P_0 = P_0 e^{5k} \Rightarrow e^{5k} = 2 \Rightarrow 5k = \ln 2 = 0.69315 \Rightarrow k = 0.69315/5 = 0.13863 ).
π Definition β Half-life: The time it takes for one-half of the atoms in an initial amount ( A_0 ) of a radioactive substance to disintegrate or transmute into atoms of another element. It measures the stability of a radioactive substance; the longer the half-life, the more stable it is. π Formula: ( kT = -\ln 2 ) or ( k = -\frac{\ln 2}{T} ) β The decay constant ( k ) is the negative natural log of 2 divided by the half-life ( T ). The initial amount is given by ( A_0 = 30e^{-30k} ).
π Example 1 (Radioactive Decay): A radioactive isotope has a half-life of 16 days. We have 30 g at the end of 30 days. How much radioisotope was initially present?
- Model: ( \frac{dA}{dt} = kA, A(30) = 30 ). Solution: ( A(t) = A_0 e^{kt} ).
- Find ( k ) using half-life: ( T=16 ). Formula: ( k = -\frac{\ln 2}{T} = -\frac{\ln 2}{16} ).
- Use condition ( A(30)=30 ): ( 30 = A_0 e^{30k} = A_0 e^{30 (-\ln 2/16)} ).
- Solve for initial amount: ( A_0 = 30 e^{-30k} = 30 e^{\frac{30 \ln 2}{16}} \approx 110.04 ) g.
π Example 2 (Finding Half-life): A breeder reactor converts uranium into plutonium 239. After 15 years, 0.043% of the initial amount ( A_0 ) of the plutonium has disintegrated. Find the half-life.
- Model: ( \frac{dA}{dt} = kA, A(0) = A_0, A(15) = (0.99957)A_0 ).
- Solution: ( A(t) = A_0 e^{kt} ).
- Find ( k ) using ( A(15) ): ( A_0 e^{15k} = (0.99957)A_0 \Rightarrow e^{15k} = 0.99957 ).
- Solve for ( k ): ( 15k = \ln(0.99957) \Rightarrow k = \frac{\ln(0.99957)}{15} \approx -0.00002867 ).
- Find half-life ( T ) using formula: ( T = \frac{\ln 2}{-\ln 2/k} ), so ( T = -\frac{\ln 2}{k} = -\frac{\ln 2}{-0.00002867} \approx 24,180 ) years.
11.1 Newton's Law of Cooling
From experimental observations, it is known that the temperature ( T(t) ) of an object changes at a rate proportional to the difference between the temperature in the body and the temperature ( T_m ) of the surrounding environment. This is what is known as Newton's law of cooling.
The differential equation is ( \frac{dT}{dt} = k(T - T_m) ), where ( k ) is a constant of proportionality. If the initial temperature of the cooling body is ( T_0 ), we obtain the initial value problem ( \frac{dT}{dt} = k(T - T_m), T(0) = T_0 ). The problem is linear as well as separable. Separating the variables and integrating, we obtain: ( \int \frac{dT}{T - T_m} = \int k dt \Rightarrow \ln |T - T_m| = kt + C_1 ). This means that ( T - T_m = e^{kt + C_1} \Rightarrow T(t) = T_m + C e^{kt} ), where ( C = e^{C_1} ). Now applying the initial condition ( T(0) = T_0 ), we see that ( C = T_0 - T_m ). Thus the solution of the initial value problem is given by: ( T(t) = T_m + (T_0 - T_m) e^{kt} ). Hence, if temperatures at times ( t_1 ) and ( t_2 ) are known, then we have: ( T(t_1) - T_m = (T_0 - T_m) e^{kt_1} ) and ( T(t_2) - T_m = (T_0 - T_m) e^{kt_2} ). π‘ Why this matters: This law governs how a cup of coffee cools or how a body's temperature changes in a forensic investigation.
π Definition β Newton's Law of Cooling: The rate of change of an object's temperature is proportional to the difference between its own temperature and the temperature of the surrounding environment. π Formula: ( T(t) = T_m + (T_0 - T_m) e^{kt} ) β The temperature at time ( t ) equals the ambient temperature plus the initial temperature difference multiplied by an exponential decay factor.
β Key Takeaways
A student must remember that the homogeneous differential equation from the previous lecture yields a family of circles and the x-axis as its solution. The exponential model ( P(t) = P_0 e^{kt} ) is the core of population dynamics and radioactive decay, which are both governed by the same differential equation, ( dA/dt = kA ). The sign of ( k ) determines growth (positive) or decay (negative). The half-life ( T ) of a substance is linked to its decay constant by the relation ( kT = -\ln 2 ). Finally, Newton's Law of Cooling, while also an exponential function, has a different formula, ( T(t) = T_m + (T_0 - T_m)e^{kt} ), where the rate of change depends on the temperature difference, not the absolute temperature.
π§ Quick Revision Questions
- What is the general solution form for both exponential population growth and radioactive decay?
- A population doubles in 10 years. What is the exact value of the growth constant ( k )?
- What is the mathematical relationship between the half-life ( T ) of a radioactive substance and its decay constant ( k )?
- In Newton's Law of Cooling, what does the constant ( T_m ) physically represent?
- If the object's initial temperature is lower than ( T_m ), will the constant of proportionality ( k ) be positive or negative in the cooling law equation?
π Lecture 13 β Higher Order Linear Differential Equations
π Overview: This lecture introduces higher-order linear differential equations, beginning with their general form and classification as homogeneous or non-homogeneous. It then formally defines the initial-value problem for these equations, laying the theoretical groundwork for solving them in subsequent lectures.
ποΈ Topics Covered
The lecture covers the preliminary theory of higher-order linear differential equations, including their general form with variable and constant coefficients, the distinction between homogeneous and non-homogeneous equations, and the formal definition and structure of the initial-value problem for an nth-order equation, with a specific look at the second-order case.
π Lecture Summary
13.1 Preliminary theory
A linear differential equation of order n can have variable coefficients and is written in the form: (a_n(x) y^{(n)} + a_{n-1}(x) y^{(n-1)} + \dots + a_1(x) y' + a_0(x) y = g(x)), where (a_n(x) \neq 0). However, this course will first focus on equations with constant coefficients, where (a_0, a_1, β¦, a_n) are real constants. When (g(x) \neq 0), the equation is called a non-homogeneous differential equation. If (g(x) = 0), the equation becomes (a_n y^{(n)} + a_{n-1} y^{(n-1)} + \dots + a_1 y' + a_0 y = 0), which is known as the associated homogeneous differential equation.
13.2 Initial-Value Problem
For a linear nth-order differential equation, an initial-value problem (IVP) is the problem of solving the differential equation subject to initial conditions at a single point (x_0). These conditions specify the value of the function and its first (n-1) derivatives at (x_0): (y(x_0) = y_0), (y'(x_0) = y'_0), β¦, (y^{(n-1)}(x_0) = y^{(n-1)}_0), where (y_0, y'_0, β¦, y^{(n-1)}_0) are arbitrary constants. For the specific case of (n=2), the second-order initial-value problem is to solve: (a_2(x) y'' + a_1(x) y' + a_0(x) y = g(x)), subject to: (y(x_0) = y_0), (y'(x_0) = y'_0).
π‘ Why this matters: This formal definition of the IVP is essential because it provides the structure for finding unique solutions to differential equations that model physical systems, where we know the system's state and its rate of change at a starting moment.
β Key Takeaways
A student must remember that a linear nth-order differential equation has a specific form with variable coefficients, but the focus in this course is on equations with constant coefficients. The critical distinction between a non-homogeneous equation and its associated homogeneous equation is fundamental. The initial-value problem (IVP) is a core concept that involves solving a differential equation subject to initial conditions, which specify the function and its first n-1 derivatives at a single point. For an nth-order equation, n initial conditions are required. The specific form of the second-order IVP should be memorized as a common case.
π§ Quick Revision Questions
- What is the general form of an nth-order linear differential equation with variable coefficients?
- What distinguishes a non-homogeneous linear differential equation from its associated homogeneous equation?
- What are the key components of an initial-value problem for a linear nth-order differential equation?
- How many initial conditions are required to define an initial-value problem for a second-order differential equation?
- In the notation for an initial-value problem, what is the variable (x_0) and what do the symbols (y_0) and (y'_0) represent?