1
First-order Differential Equations
Consider cases of the form y 0 (t) = F (t, y)
1.1
Linear ODEs
If we have a differential equation of the form:
y 0 + p(t)y = g(t)
Then we can solve this by computing the integrating factor = e
Z
y(t) = ( g(t))
1.2
p(t)dt
. Then,
Exact ODEs
If we have a differential equation of the form:
dy
=0
dt
R
R
and My = Nt , then the equation is exact, and we can compute a function = M dt + h(y) = N dy + h(t).
We then differentiate to solve for the unknown function.
M +N
The family of solutions is then contained in the level curves of = K.
1.3
Homogenous ODEs
If we have a differential equation y 0 (t) = F (t, y) such that F (at, ay) = F (t, y), then it is homogenous. This
is equivalent to being able to write F (t, y) as F ( yt ), i.e a function of yt .
In this case, we introduce a new variable z = yt y = zt y 0 = z + tz 0 . Then, we solve this new ODE
for z and express the result in terms of y, t at the end.
Systems of ODEs
Here we consider systems of ODEs of the form:
y 0 = Ay
where A is a 2 2 or 3 3 matrix. We proceed by computing the eigenvalues of A, which are the solutions
of the following quadratic equation:
2 (tr A) + det A
2.1
Distinct Real Eigenvalues
If A has two real eigenvalues 1 , 2 , then proceed by computing the corresponding eigenvectors v1 , v2 . Then,
the general solution to the system is:
y = C1 e1 t v1 + C2 e2 t v2
Classifying the critical point:
If 1 > 2 > 0, the origin is a source.
1
If 2 < 1 < 0, the origin is a sink.
If 2 < 0 < 1 , the origin is a saddle point
2.2
Complex Eigenvalues
Here we consider the case when the eigenvalues of A are i.
Proceed by finding an eigenvector v for either of the eigenvalues. Then, consider:
e+i v = e ei v
and split ei using Eulers formula, which is:
ei = cos(t) + i sin(t)
.
Then, you will be able to split e ei v = u(t) + iv(t), which are its real and imaginary parts respectively.
The general solution to the problem will then be:
y = C1 u(t) + C2 v(t)
.
Classifying the critical point:
If > 0, the origin is a spiral source.
If < 0, the origin is a spiral sink.
If = 0, the origin is a spiral center and is periodic with period
2.3
2
|| .
Repeated Eigenvalues
We may find ourselves in a situation where 1 = 2 .
In this case, the general solution is:
y = (1 )et y0 + tet Ay0
where y0 is the given initial condition. If we are not given an initial
condition, we can find a fundamental
set of solutions by plugging in the initial conditions y0 = 10 , 01 .
2.4
The Wronkskian Determinant
The wronkskian determinant of a set of solutions is the determinant of the 2 2 matrix that has them as
columns. If this determinant is never equal to 0, the two solutions form a fundamental set of solutions.
To find the Wronkskian at an arbitrary general time: W (t) = e(tr A)t W (0).
2.5
Trace-Determinant Classification
It is not necessary to compute the eigenvalues to determine the type of critical point the origin is - we just
need to know the determinant and trace:
If det A < 0, the origin is a saddle point.
If det A > 0
If tr A > 0, the origin is a source.
If tr A < 0, the origin is a sink.
If tr A = 0, the origin is a center.
2
2.6
Variation of Parameters
This section deals with solving inhomogenous equations of the form:
y 0 = Ay + g(t)
We can solve problems of this nature by first finding a fundamental set of solutions y1 , y2 to the homogenous
problem y 0 = Ay. Then, observe that the solution to the inhomogenous equation must be of the form:
y(t) = u1 (t)y1 + u2 (t)y2
.
We then solve for u01 , u02 with the following equation:
u01 y1 + u02 y2 = g(t)
Then, integrate to solve for u1 , u2 and the solution is given.
Second Order ODEs
Second order ODEs are of the form ay 00 + by 0 + cy = g(t)
3.1
Homogenous Second Order ODEs
To solve an second order ODE of the form
ay 00 + by 0 + cy = 0
we need to solve the characteristic equation a2 + b + c = 0 to get the roots 1 , 2
If 1 , 2 are distinct and real, the general solution is y = C1 e1 t + C2 e2 t
If 1 = 2 are repeated roots, the general solution is y = C1 et + C2 tet
If = i are complex roots, the general solution is y = C1 et cos(t) + C2 et sin(t)
3.2
Nonhomogenous ODEs - Method of Underdetermined Coefficients
If we want to solve a nonhomogenous second order ODE, such as ay 00 + by 0 + cy = g(t), one needs.
3.3
The Laplace Transform
The Laplace Transform of some common functions:
L(1) =
1
s
L(tn ) =
n!
sn+1
L(ect ) =
1
sc
L(tn ect ) =
n!
(sc)n+1
L(cos(bt)) =
s
s2 +b2
L(sin(bt)) =
b
s2 +b2
Non-linear Systems of ODEs
4.1
Hamiltonian Systems
Hamiltonian systems are a special case of systems which contain an embedded conservation law. Consider
the following system:
x0 = f (x, y) and y 0 = g(x, y)
.
The system is Hamiltonian if
F
x
G
y
= 0.
If this is the case, then there is a constant function called the Hamiltonian H(x, y), which has the following
properties:
H
H
= G and
=F
x
y
.
From here, we can solve for H similar to the manner in which you determine in the situation of exact
equations.
4.2
Determining the Stability of Critical Points
We can determine the stability of critical points of nonlinear systems by linearizing the system and applying
the trace-determinant criterion. We go from the following system:
x0 = f (x, y) and y 0 = g(x, y)
.
to
z0 =
F
x
G
x
F
y
G
y
!
z
If the linearized systems origin is a sink, then the non-linear system is asymptotically stable. If it is a source
or a saddle, then it is unstable.
