FP2.4 - First order differential equations

Syllabus
2019
Topic
Level
A2

Learning objectives

Solve and sketch separable first-order equations

A first-order equation is separable when it can be written as dydx=g(x)h(y),\frac{dy}{dx}=g(x)h(y), so every factor involving yy can be placed with dydy and every factor involving xx with dxdx. The constant of integration produces a family of solution curves; an initial condition selects one member.

Step Reason
Form the equation Translate a rate statement into dy/dxdy/dx and define every variable and constant.
Record equilibrium solutions If h(y)=0h(y)=0, constant curves may solve the equation and would be lost by division.
Separate and integrate Write 1h(y)dy=g(x)dx\dfrac{1}{h(y)}\,dy=g(x)\,dx, then include one arbitrary constant.
Apply the condition Substitute the given (x,y)(x,y) pair to obtain the particular solution.
Analyse before sketching Find the domain, asymptotes and stationary points; use the sign of dy/dxdy/dx for increasing or decreasing branches.

For dydx=x(1+y),y(0)=0,\frac{dy}{dx}=x(1+y),\qquad y(0)=0, first record the equilibrium solution y=1y=-1. For the other solutions,11+ydy=xdx,\int\frac{1}{1+y}\,dy=\int x\,dx,so ln1+y=x22+Cy=Aex2/21.\ln|1+y|=\frac{x^2}{2}+C\quad\Longrightarrow\quad y=Ae^{x^2/2}-1.The condition y(0)=0y(0)=0 gives A=1A=1, hence y=ex2/21y=e^{x^2/2}-1.

For the full family y=Aex2/21y=Ae^{x^2/2}-1, every curve is symmetric about the yy-axis and has dy/dx=0dy/dx=0 at x=0x=0. Curves with A>0A>0 lie above the equilibrium and have a minimum at x=0x=0; curves with A<0A<0 lie below it and have a maximum there. This qualitative information should agree with any plotted member.

Do not divide by a function of yy before checking where it is zero, because this can discard equilibrium solutions. Keep the integration constant until an initial or boundary condition is applied, and state the interval on which logarithms, denominators and the explicit solution are defined. A sketch must follow the solved curve and the differential equation's gradient signs, not a generic shape.

Solve a linear first-order equation with an integrating factor

A first-order linear equation must first be written in the standard form dydx+P(x)y=Q(x).\frac{dy}{dx}+P(x)y=Q(x). Its integrating factor is μ(x)=eP(x)dx.\mu(x)=e^{\int P(x)\,dx}. Because μ=Pμ\mu'=P\mu, multiplying by μ\mu turns the entire left side into the product derivative ddx(μy)\dfrac{d}{dx}(\mu y).

ddx(μy)=μQy=μ(x)Q(x)dx+Cμ(x)\frac{d}{dx}(\mu y)=\mu Q\quad\Longrightarrow\quad y=\frac{\int \mu(x)Q(x)\,dx+C}{\mu(x)}

Step Check
Normalise Divide by the coefficient of dy/dxdy/dx; identify PP and QQ only afterwards.
Build μ\mu Integrate PP with the correct sign; a non-zero constant factor in μ\mu is immaterial.
Multiply every term Verify that the left side equals (μy)(\mu y)'.
Integrate once Integrate μQ\mu Q, add CC, then divide by μ\mu.
Particularise and verify Apply the condition, then substitute or differentiate to check the result.

For dydx2y=x,\frac{dy}{dx}-2y=x, we have P=2P=-2 and μ=e2x\mu=e^{-2x}. Thus (e2xy)=xe2x.(e^{-2x}y)'=xe^{-2x}.Integration by parts gives e2xy=e2x(x2+14)+C,e^{-2x}y=-e^{-2x}\left(\frac{x}{2}+\frac14\right)+C, so y=x214+Ce2x.y=-\frac{x}{2}-\frac14+Ce^{2x}. If y(0)=34y(0)=\tfrac34, then C=1C=1.

The term containing CC is the homogeneous part of the family, while the remaining terms form one particular response to Q(x)Q(x). Their sum is still a solution because the equation is linear.

The integrating-factor formula applies after the equation is in standard linear form; PP and QQ may be functions of xx but not of yy. Multiplying only some terms, using ePe^{P} instead of ePdxe^{\int P\,dx}, or forgetting to divide the final expression by μ\mu breaks the product-derivative identity. Preserve the stated xx-domain when division or logarithms are involved.

Use a given substitution to reduce a first-order equation

A given substitution is useful when it converts a non-standard first-order equation into either a separable equation or a linear equation. The essential move is to differentiate the substitution with the chain rule, replace every occurrence of the old dependent variable consistently, and solve the resulting equation in the new variable.

Given form Differentiate as
v=yg(x)v=y-g(x) dvdx=dydxg(x)\dfrac{dv}{dx}=\dfrac{dy}{dx}-g'(x)
v=ynv=y^n dvdx=nyn1dydx\dfrac{dv}{dx}=ny^{n-1}\dfrac{dy}{dx}
v=ynv=y^{-n} dvdx=nyn1dydx\dfrac{dv}{dx}=-ny^{-n-1}\dfrac{dy}{dx}
y=F(v)y=F(v) dydx=F(v)dvdx\dfrac{dy}{dx}=F'(v)\dfrac{dv}{dx}

After differentiating, substitute into the original equation and simplify until only xx, vv and dv/dxdv/dx remain. Identify the transformed equation as separable or linear, solve it by the corresponding earlier method, restore the integration constant, and finally replace vv to express the result in the requested variable and form.

Consider dydx+2y=xy3\frac{dy}{dx}+2y=xy^3 with the given substitution v=y2v=y^{-2}. Since dvdx=2y3dydx,\frac{dv}{dx}=-2y^{-3}\frac{dy}{dx}, multiplying the original equation by 2y3-2y^{-3} gives dvdx4v=2x.\frac{dv}{dx}-4v=-2x.This is linear. With integrating factor e4xe^{-4x},ve4x=2xe4xdx=e4x(x2+18)+C,ve^{-4x}=\int-2xe^{-4x}\,dx=e^{-4x}\left(\frac{x}{2}+\frac18\right)+C,so v=x2+18+Ce4x,y2=1x2+18+Ce4x.v=\frac{x}{2}+\frac18+Ce^{4x},\qquad y^2=\frac{1}{\frac{x}{2}+\frac18+Ce^{4x}}.

The substitution may impose restrictions. Here v=y2v=y^{-2} assumes y0y\ne0, so the original equation must be checked separately: y=0y=0 is also a solution and was excluded during the transformation. Do not stop at the equation in vv unless that is what was requested, and do not confuse this first-order reduction with the separate second-order workflow in FP2.5.