Review - Solutions of Ordinary Differential Equations

Author: John M. Cimbala, Penn State University

Latest revision: 06 September 2007

First-Order, Separable, Ordinary Differential Equations

Example:
y = y(t)

dy
=C
dt

Example:
y = y(t)

ty

dy
=C
dt

Separate variables: ydy = Cdt

Solve: y 2 / 2 = Ct + C1

Separate variables: ydy =

2
Solve: y / 2 = C ln ( t ) + C1

C
dt
t

First-Order, Linear, Non-Homogeneous, Ordinary Differential Equations

Example:
y = y(x)

dy
+ P ( x) y = Q ( x)
dx

( Pdx )

Multiply each term by an integrating factor, v = exp

Solve: This leads to an exact differential on the LHS, which can be solved
easily.

Higher-Order, Linear, Homogeneous, Ordinary Differential Equations with Constant Coefficients

Example:
y = y(x)

d3y
dx 3

d2y
dx 2

dy
+ 6y = 0
dx

Rewrite in terms of a linear operator, D n y

(D3 - 4D2 + D + 6) y = 0
Factor the LHS: Here, (D + 1) (D - 2) (D - 3) y = 0
Find the roots of the LHS: Here the roots are -1, 2, and 3
Solve: y = C1e x + C2 e 2 x + C3e3 x

Example:
y = y(x)

d y

Example:
y = y(x)

dx 2

a2 y = 0

d y
+ a2y = 0
dx 2

dn y

dx n

: Here,

Rewrite in terms of the linear operator: Here, (D2 - a) y = 0

Factor the LHS: Here, (D - a) (D + a) y = 0
Find the roots of the LHS: Here the roots are a and -a
Solve: y = C1e ax + C2 e ax or y = C3 cosh(ax) + C4 sinh(ax)
Rewrite in terms of the linear operator: Here, (D2 + a) y = 0
Factor the LHS: Here, (D - ia) (D + ia) y = 0
Find the roots of the LHS: Here the roots are ia and -ia
Solve: y = C1eiax + C2 e iax or y = C3 cos(ax) + C4 sin(ax)

Higher-Order, Nonlinear, Non-Homogeneous, Ordinary Differential Equations

Example:
y = y(x)

d3y
dx

+ P ( x)

d2y

dy
+ = Q ( x)
2
dx
dx

Define N unknowns, where N is the highest order derivative of the

original equation: Here, N = 3, and define

d2y

dy
, F3 y
dx
dx
Split equation into N first-order equations, one for each F:
F1

, F2

dF3 dy
dF1 d 3 y
dF2 d 2 y
d 2 y dy
= 3 = Q( x) P ( x) 2 ,
=
= 2 ,
dx dx
dx dx
dx
dx
dx
dx
Define derivative array in standard Runge-Kutta form, in terms of the
N unknowns, F1, F2, ..., FN:
dF
dF
dF
D1 1 = Q( x) P( x) F1 F2 2 , D2 2 = F1 , D3 3 = F2
dx
dx
dx
Solve simultaneously for F1, F2, ..., FN using the Runge-Kutta
numerical technique and the boundary conditions.