1.08k, 1.08l - Separable differential equations and their solutions

1.08k, 1.08l - Separable differential equations and their solutions

Find general and particular solutions of first-order differential equations by direct integration or separation of variables. You will keep constant solutions, select valid branches and interpret rates, limits and restrictions in a motion model.

Solutions and initial information

Can a falling object keep speeding up without ever reaching 12 m s−112\,\mathrm{m\,s^{-1}}? A supplied model gives its downward velocity vv by dv/dt=6−v/2\mathrm dv/\mathrm dt=6-v/2, starting from rest. At v=0v=0, its acceleration is 66; at v=6v=6, it is 33; at v=9v=9, it is only 1.51.5, all in m s−2\mathrm{m\,s^{-2}}. Solving the equation will tell us whether the speed reaches 1212 and when it reaches halfway.

You already know

Integration reverses differentiation. An indefinite integral needs an arbitrary constant. The chain rule differentiates a composite function, and ∫1/x dx=ln⁡∣x∣+C\int 1/x\,\mathrm dx=\ln|x|+C for x≠0x\ne0.

A solution must satisfy the rate rule throughout its interval, not merely pass through one point.

General and particular solutions

A solution is a differentiable function satisfying the differential equation. A general solution contains an arbitrary constant describing a family of solutions. An initial condition, such as y(−1)=1y(-1)=1, can select a particular solution. For example, every y=x2+Cy=x^2+C solves dy/dx=2x\mathrm dy/\mathrm dx=2x, but only C=3C=3 gives y(0)=3y(0)=3.

If the derivative depends only on the independent variable, integrate directly. This solution method is not in the formula sheet: learn it.

Worked example 1: a point selects a curve

Question. Solve dy/dx=3x2+2\mathrm dy/\mathrm dx=3x^2+2, given y(−1)=1y(-1)=1.

Plan. Integrate with a constant, then substitute both coordinates of the initial point.

y=x3+2x+Cintegrate each term1=(−1)3+2(−1)+Cuse y(−1)=1C=4solve for the constant\begin{aligned} y&=x^3+2x+C && \text{integrate each term} \\ 1&=(-1)^3+2(-1)+C && \text{use }y(-1)=1 \\ C&=4 && \text{solve for the constant} \end{aligned}

Why this step? The given ordinate is 11, but the other terms at x=−1x=-1 contribute −3-3, so CC is not 11.

Answer. y=x3+2x+4y=x^3+2x+4, for real xx.

Check. Differentiating gives 3x2+23x^2+2, and y(−1)=−1−2+4=1y(-1)=-1-2+4=1.

Change only the input of the initial point from −1-1 to 11. The derivative stays the same, so the curves differ only by a vertical translation.

Worked example 2: a different initial input

Question. Solve dy/dx=3x2+2\mathrm dy/\mathrm dx=3x^2+2, given y(1)=1y(1)=1.

Plan. Use the same general solution and fit its constant to the new point.

y=x3+2x+Cgeneral solution1=1+2+Csubstitute x=1C=−2determine the translation\begin{aligned} y&=x^3+2x+C && \text{general solution} \\ 1&=1+2+C && \text{substitute }x=1 \\ C&=-2 && \text{determine the translation} \end{aligned}

Why this step? A point condition selects the member of the family; it does not change the differential equation.

Answer. y=x3+2x−2y=x^3+2x-2.

Check. The derivative is still 3x2+23x^2+2, and y(1)=1+2−2=1y(1)=1+2-2=1.

A domain can split a family into separate intervals. For dy/dx=1/(x−2)\mathrm dy/\mathrm dx=1/(x-2), direct integration gives y=ln⁡∣x−2∣+Cy=\ln|x-2|+C. The equation is undefined at 22, so a condition at x=3x=3 selects an interval on the right, never a solution passing through the pole.

Watch out

Wrong: take the initial ordinate as CC without substitution. Right: substitute the entire point into the integrated equation. Then check both the derivative and the point condition.

Move the initial input once more in the next problem.

Separating variables before integration

When the right-hand side involves both variables, first look for a product of a function of xx and a function of yy. Factorisation may be the step that reveals it.

Separation of variables

If

dydx=g(x)h(y),\frac{\mathrm dy}{\mathrm dx}=g(x)h(y),

then, where h(y)≠0h(y)\ne0, separate and integrate:

∫1h(y) dy=∫g(x) dx.\int\frac1{h(y)}\,\mathrm dy=\int g(x)\,\mathrm dx.

Check constant solutions from h(y)=0h(y)=0 separately, and restrict the solution to an interval where the equation is defined.

Not in the formula sheet: learn it.

The notation is backed by the chain rule: differentiating the left integral with respect to xx produces (1/h(y))dy/dx(1/h(y))\mathrm dy/\mathrm dx. It is not permission to treat a varying yy as constant in an integral with respect to xx.

Worked example 3: factor before separating

Question. Solve dy/dx=xy+y\mathrm dy/\mathrm dx=xy+y, given y(0)=2y(0)=2.

Plan. Factor out yy, separate the variables, then use the initial condition after integrating.

dydx=(x+1)ycommon factor∫1y dy=∫(x+1) dxy≠0ln⁡∣y∣=x22+x+Cintegratey=Aex2/2+xabsorb the sign into A2=Ae0⟹A=2initial condition\begin{aligned} \frac{\mathrm dy}{\mathrm dx}&=(x+1)y && \text{common factor} \\ \int\frac1y\,\mathrm dy&=\int(x+1)\,\mathrm dx && y\ne0 \\ \ln|y|&=\frac{x^2}{2}+x+C && \text{integrate} \\ y&=Ae^{x^2/2+x} && \text{absorb the sign into }A \\ 2&=Ae^0\quad\Longrightarrow\quad A=2 && \text{initial condition} \end{aligned}

Why this step? Exponentiating converts an additive logarithmic constant into a multiplicative constant. The positive initial value selects a positive branch.

Answer. y=2ex2/2+xy=2e^{x^2/2+x}, for real xx. The constant solution y=0y=0, excluded during division, does not meet this initial condition.

Check. Differentiation gives y′=(x+1)yy'=(x+1)y, and y(0)=2y(0)=2.

Change only the dependent-variable factor from yy to y−1y-1. Separate the whole factor; the quantity inside the logarithm changes with it.

Worked example 4: a shifted dependent variable

Question. Solve dy/dx=xy+y−x−1\mathrm dy/\mathrm dx=xy+y-x-1, given y(0)=2y(0)=2.

Plan. Factor into (x+1)(y−1)(x+1)(y-1), then integrate the reciprocal of y−1y-1.

dydx=(x+1)(y−1)group and factor∫1y−1 dy=∫(x+1) dxy≠1ln⁡∣y−1∣=x22+x+Cintegratey−1=Aex2/2+xexponentiate2−1=A⟹A=1fit the point\begin{aligned} \frac{\mathrm dy}{\mathrm dx}&=(x+1)(y-1) && \text{group and factor} \\ \int\frac1{y-1}\,\mathrm dy&=\int(x+1)\,\mathrm dx && y\ne1 \\ \ln|y-1|&=\frac{x^2}{2}+x+C && \text{integrate} \\ y-1&=Ae^{x^2/2+x} && \text{exponentiate} \\ 2-1&=A\quad\Longrightarrow\quad A=1 && \text{fit the point} \end{aligned}

Why this step? The common factor is y−1y-1, so dividing by yy would not separate the equation.

Answer. y=1+ex2/2+xy=1+e^{x^2/2+x}, for real xx. The separate equilibrium y=1y=1 fails the initial condition.

Check. y′=(x+1)ex2/2+x=(x+1)(y−1)y'=(x+1)e^{x^2/2+x}=(x+1)(y-1), and y(0)=2y(0)=2.

The same calculation solves proportional decay. For P′=−0.2PP'=-0.2P, P(0)=50P(0)=50, separation gives ln⁡P=−0.2t+C\ln P=-0.2t+C, then P=50e−0.2tP=50e^{-0.2t}. Its derivative is negative, but its magnitude decreases as the positive quantity gets smaller. These exponential solutions are derived here; they are not printed in the formula sheet.

Watch out

Wrong: y=ex2/2+x+Cy=e^{x^2/2+x}+C after ln⁡∣y∣=x2/2+x+C\ln|y|=x^2/2+x+C. Right: y=Aex2/2+xy=Ae^{x^2/2+x}. A real signed AA includes the branch sign; check A=0A=0 in the equation before including it.

For the next shifted equation, change the scale of the right-hand side and the initial value. Show the separation rather than only writing a remembered exponential.

Recovering equilibria and choosing a valid branch

Division can erase a perfectly good constant solution. Record the zero factors first, and use the initial value to choose the relevant branch afterwards.

Equilibrium solution

An equilibrium solution is a constant function satisfying the differential equation. For y′=y(2−y)y'=y(2-y), both y=0y=0 and y=2y=2 have derivative zero and make the right-hand side zero. A constant such as y=1y=1 is not an equilibrium: its derivative is zero but the right-hand side is 11.

Partial fractions from your earlier integration work handle a product of two linear factors. Keep the minus sign from integrating 1/(2−y)1/(2-y).

Worked example 5: keep both constant solutions

Question. Identify the equilibria of y′=y(2−y)y'=y(2-y), then solve with y(0)=1y(0)=1.

Plan. Check the zero factors, separate the nonconstant solutions and use partial fractions.

y(2−y)=0⟹y=0 or y=2equilibria1y(2−y)=12y+12(2−y)partial fractions12ln⁡∣y∣−12ln⁡∣2−y∣=x+Cseparate and integratey2−y=Ke2xcombine logarithms1=K⟹y=21+e−2xuse y(0)=1\begin{aligned} y(2-y)&=0\quad\Longrightarrow\quad y=0\text{ or }y=2 && \text{equilibria} \\ \frac1{y(2-y)}&=\frac1{2y}+\frac1{2(2-y)} && \text{partial fractions} \\ \frac12\ln|y|-\frac12\ln|2-y|&=x+C && \text{separate and integrate} \\ \frac{y}{2-y}&=Ke^{2x} && \text{combine logarithms} \\ 1&=K\quad\Longrightarrow\quad y=\frac2{1+e^{-2x}} && \text{use }y(0)=1 \end{aligned}

Why this step? Division excluded both equilibria. The logarithmic relation describes only the other solutions, so it cannot by itself be the complete answer about equilibria.

Answer. The equilibria are 00 and 22. The particular solution is y=2/(1+e−2x)y=2/(1+e^{-2x}), for real xx.

Check. With z=e−2x>0z=e^{-2x}>0, y′=4z/(1+z)2=y(2−y)y'=4z/(1+z)^2=y(2-y), and y(0)=1y(0)=1. Also 0<y<20<y<2.

Continue this lesson.

You’re previewing NovaLearn without an account. Create a free account to read the rest.

Create free account