Second-order differential equations reducible to first order
The general second-order differential equation has the form:
a(x,y)dx2d2y+b(x,y)dxdy+c(x,y)y=g(x,y)(7.15)
If a, b, and c are functions of x only, the equation is linear. If a, b, and c are constants, the linear equation becomes:
adx2d2y+bdxdy+cy=g(x)(7.16)
When c=0, equation (7.16) can be reduced to a first-order differential equation.
Case 1: Dependent variable y is missing
If the equation is of the form F(x,y′,y′′)=0, we substitute P=y′, then y′′=dxdP. This transforms the second-order equation into a first-order equation in terms of x and P.
Case 2: Independent variable x is missing
If the equation is of the form F(y,y′,y′′)=0, we use the substitution P=y′, then y′′=dxdP=dydPdxdy=PdydP. This transforms the second-order equation into a first-order equation in terms of y and P.
Example 1
Solve each of the following differential equations:
(a) xy′′=y′
(b) yy′′+(y′)2=y(y′)3
Solution:
(a) xy′′=y′. Let P=y′, so y′′=dxdP. Substituting gives:
xdxdP=P
PdP=xdx
∫PdP=∫xdx
ln∣P∣=ln∣x∣+C1
P=C1x
Since P=y′, we have dxdy=C1x
dy=C1xdx
∫dy=∫C1xdx
y=21C1x2+C2,ory=Ax2+B
(b) yy′′+(y′)2=y(y′)3. Let P=y′, so y′′=PdydP. Substituting gives:
yPdydP+P2=yP3
Divide by yP:
dydP+yP=P2
This is a Bernoulli equation. Let v=P−1, so P=v−1 and dydP=−v−2dydv. Substituting gives:
−v−2dydv+yv−1=v−2
dydv−yv=−1
Integrating factor: e∫−y1dy=e−ln∣y∣=y1
y1dydv−y2v=−y1
dyd(yv)=−y1
yv=−ln∣y∣+C1
v=−yln∣y∣+C1y
P=−yln∣y∣+C1y1
dxdy=y(−ln∣y∣+C1)1
∫y(−ln∣y∣+C1)dy=∫dx
This integral is complex and won't be solved here. The original text also did not solve it.
Example 2
Find the particular solution in the following boundary-value problem:
(a) yy′′+5(y′)2=25y; y(0)=34 and y′(0)=1
Solution:
yy′′+5(y′)2=25y. Let P=y′, so y′′=PdydP. Substituting gives:
yPdydP+5P2=25y
PdydP+5P=25
dydP=P25−5P
25−5PPdP=dy
∫25−5PPdP=∫dy
Solving the integral on the left side (using substitution or partial fractions) gives:
−51P−251ln∣25−5P∣=y+C1
This is hard to solve for P. The original text also made an error here. It appears the example was not well chosen.
Solution of a non-homogeneous second-order differential equation
Consider the non-homogeneous second-order differential equation:
a2dx2d2y+a1dxdy+a0y=f(x)(7.17)
where a2, a1, and a0 are constants, and f(x)=0. The general solution is y=yc+yp, where:
- yc is the complementary function: the general solution of the corresponding homogeneous equation (a2y′′+a1y′+a0y=0).
- yp is a particular integral: a solution of the non-homogeneous equation containing no arbitrary constants.
The form of yp depends on f(x). Crucially, yp must be linearly independent of any term in yc.
Here's how to choose yp, based on f(x):
- f(x)=keax (k and a are constants):
- If eax is not a term in yc: yp=Aeax
- If eax is a term in yc: yp=Axeax
- If both eax and xeax are in yc: yp=Ax2eax
- f(x) is a polynomial of degree n: f(x)=Pn(x)=knxn+kn−1xn−1+…+k1x+k0 (where ki are constants) Assume: yp=Axn+Bxn−1+…+Cx+D
- f(x)=ksin(βx) or f(x)=kcos(βx) (or a linear combination): Assume: yp=Asin(βx)+Bcos(βx)
- f(x)=Pn(x)eax (a polynomial times an exponential): Assume: yp=eax(Axn+Bxn−1+…+Cx+D)
- f(x)=Pn(x)sin(βx) or f(x)=Pn(x)cos(βx) (a polynomial times a trigonometric function): Assume: yp=(Axn+Bxn−1+…+Cx+D)sin(βx)+(Exn+Fxn−1+…+Gx+H)cos(βx)
- f(x)=eaxsin(βx) or f(x)=eaxcos(βx) (an exponential times a trigonometric function): Assume: yp=eax(Asin(βx)+Bcos(βx))
Example 1
Find the general solution of each of the following differential equations:
(a) y′′−2y′−3y=3ex
(b) y′′−2y′−3y=3e3x
Solution:
(a) y′′−2y′−3y=3ex
The characteristic equation for the homogeneous part is λ2−2λ−3=0, which factors to (λ−3)(λ+1)=0. The roots are λ1=3 and λ2=−1. The complementary function is:
yc=Ae3x+Be−x
Since f(x)=3ex and ex is not part of yc, we assume a particular integral of the form yp=λex.
yp′=λex
yp′′=λex
Substituting into the original equation:
λex−2λex−3λex=3ex
−4λex=3ex
λ=−43
So, yp=−43ex.
The general solution is:
y=yc+yp=Ae3x+Be−x−43ex
(b) y′′−2y′−3y=3e3x
The complementary function is the same as in part (a): yc=Ae3x+Be−x.
Since f(x)=3e3x and e3x is part of yc, we assume a particular integral of the form yp=λxe3x.
yp′=λe3x+3λxe3x
yp′′=6λe3x+9λxe3x
Substituting into the original equation:
6λe3x+9λxe3x−2(λe3x+3λxe3x)−3(λxe3x)=3e3x
4λe3x=3e3x
λ=43
So, yp=43xe3x.
The general solution is:
y=yc+yp=Ae3x+Be−x+43xe3x
Example 2
Solve the initial-value problem: y′′−2y′−3y=3t2+4t−5; y(0)=6 and y′(0)=19.
Solution:
The complementary function is the same as in Example 1(a): yc=c1e−t+c2e3t.
Since f(t)=3t2+4t−5, we assume yp=At2+Bt+C.
yp′=2At+B
yp′′=2A
Substituting into the original equation:
2A−2(2At+B)−3(At2+Bt+C)=3t2+4t−5
−3At2+(−4A−3B)t+(2A−2B−3C)=3t2+4t−5
Equating coefficients:
−3A=3⇒A=−1
−4A−3B=4⇒4−3B=4⇒B=0
2A−2B−3C=−5⇒−2−0−3C=−5⇒C=1
So, yp=−t2+1.
The general solution is y=c1e−t+c2e3t−t2+1.
Applying initial conditions:
y(0)=6: 6=c1+c2+1⇒c1+c2=5
y′(t)=−c1e−t+3c2e3t−2t
y′(0)=19: 19=−c1+3c2
Solving the system of equations gives c1=−1 and c2=6.
The particular solution is y=−e−t+6e3t−t2+1.