Differential Equations Study Sheet

Matthew Chesnes
It’s all about the Mathematics!
Kenyon College
Exam date: May 11, 2000
6:30 P.M.
1 First Order Differential Equations
• Differential equations can be used to explain and predict new facts for about everything
that changes continuously.
d2 x dx
• + a + kx = 0.
dt2 dt
• t is the independent variable, x is the dependent variable, a and k are parameters.

• The order of a differential equation is the highest deriviative in the equation.

• A differential equation is linear if it is linear in parameters such that the coefficients

on each deriviative of y term is a function of the independent variable (t).

• Solutions: Explicit → Written as a function of the independent variable. Implicit →

Written as a function of both y and t. (defines one or more explicit solutions.

1.1 Population Model

dP
• Model: = kP .
dt
dP
• Equilibrium solution occurs when = 0.
dt
• Solution: P (t) = Ae(kt) .

• If k > 0, then limt→∞ P (t) = ∞. If k < 0, then limt→∞ P (t) = 0.

• Redefine model so it doesn’t blow up to infinity.

dP P
• = kP (1 − ).
dt N
• N is the carrying capacity of the population.

1.2 Seperation of Variables Technique

dy
• = g(t)h(y).
dt
1
• dy = g(t)dt.
h(y)
• Integrate both sides and solve for y.

• You might lose the solution h(y) = 0.

2
1.3 Mixing Problems
dQ
• = Rate In - Rate Out.
dt
• Consider a tank that initally contains 50 gallons of pure water. A salt solution contain-
ing 2 pounds of salt per gallon of water is poured into the tank at a rate of 3 gal/min.
The solution leaves the tank also at 3 gal/min.

• Therefore Input = 2(lb/gal)*3(gal/min).

• Output = ?(lbs/gal)*3(gal/min).
Q(t)
• Salt in Tank = .
50
Q(t)
• Therefore output of salt = (lbs/gal)*3(gal/min).
50
dQ Q(t)
• = Rate In - Rate Out = 2 lbs/gal*3gal/min - (lbs/gal)*3(gal/min).
dt 50
3Q
• 6 lbs/min - lbs/min.
50
• Solve via seperation of Variables.

1.4 Existance and Uniqueness

dy
• Given = f (t, y). If f is continuous on some interval, then there exists at least one
dt
solution on that interval.
∂
• If both f (t, y) and f (t, y) are continuous on some interval then an initial value
∂y
problem on that interval is guaranteed to have exactly one Unique solution.

1.5 Phase Lines

• Takes all the information from a slope fields and captures it in a single vertical line.

• Draw a vertical line, label the equilibrium points, determine if the slope of y is positive
or negative between each equilibrium and label up or down arrows.

1.6 Classifying Equilibria and the Linearization Theorem

• Source: solutions tend away from an equilibrium → f 0 (yo ) > 0.

• Sink: solutions tend toward an equilibrium → f 0 (yo ) < 0.

• Node: Nither a source or a sink → f 0 (yo ) = 0 or DNE.

3
1.7 Bifurcations
• Bifurcations occur at parameters where the equilibrium profile changes.

• Draw phase lines (y) for several values of a.

1.8 Linear Differential Equations and Integrating Factors

• Properties of Linear DE: If yp and yh are both solutions to a differential equation,
(particular and homogeneous), then yp + yh is also a solution.
dy
• Using the integrating factor to solve linear differential equations such that +P (t)y =
dt
f (t).
R
• The integrating factor is therefore e( P (t)dt)
.

• Multiply both sides by the integrating factor.

P (t)dt) dy
R R R
• e( + e( P (t)dt)
P (t)y = e( P (t)dt)
f (t).
dt
• then via chain rule ...
R
d{e( P (t)dt)
y} R
• ((Integrating factor * y))= e( P (t)dt)
f (t).
dt
• Then integrate to find solution.

1.9 Integration by Parts

R R
udv = uv − vdu.

4
2 Systems
dx dy
• = ax − bxy, = −cy + dxy.
dt dt
• Equilibrium occurs when both differential equations are equal to zero.

• a and c are growth effects and b and d are interaction effects.

• To verify that x(t), y(t) is a solution to a system, take the deriviative of each and
compare them to the originial differerential equations with x and y plugged in.
d2 y dy d2 y
• Converting a second order differential equation, = y. Let v = . Thus dv = .
dt2 dt dt

2.1 Vector Notation

dx dy
• A system of the form = ax + bxy and = cy + exy can be written in vector
dt dt
notation.

•  
dx  
d  dt  ax + bxy
P(t) =  dy  = . (1)
dt cy + exy
dt

2.2 Decoupled System

dx dy
• Completely decoupled: = f (x), = g(y).
dt dt
dx dy
• Partially decoupled: = f (x), = g(x, y).
dt dt

5
3 Systems II
• Matrix form.
d
• Homogeneous = X = AX.
dt
d
• Non-homogeneous = X = AX + F.
dt
• Linearity Principal
d
• Consider X = AX,where
dt  
a b
A= . (2)
c d

• If X1 (t) and X2 (t) are solutions, then k1 X1 (t) + k2 X2 (t) is also a solution provided
X1 (t) and X2 (t) are linearly independent.

• Theorem: If A is a matrix wtih det A not equal to zero, then the only equilibrium
d
piont for the system X = AX is,
dt
 
0
. (3)
0

3.1 Straightline Solutions, Eigencool Eigenvectors and Eigenval-

ues
d
• A straightline solution to the system X = AX exists provided that,
dt
   
x x
A =λ . (4)
y y

• To determine λ, compute the det[(A - λI)] =

 
a−λ b
det = (a − λ)(e − λ) − bc = 0. (5)
c e−λ
.

• This expands to the characteristic polynomial =

λ2 − (a − d)λ + ae − bc = 0.

• Solving the characteristic polynomial provides us with the eigenvalues of A.

6
3.2 Stability
Consider a linear 2 dimensional system with two nonzero, real, distinct eigenvalues, λ1 and
λ2 .

• If both eigenvalues are positive then the origin is a source (unstable).

• If both eigenvalues are negative then the origin is a sink (stable).

• If the eigenvalues have different signs, then the origin is a saddle (unstable).

3.3 Complex Eigenvalues

• Euler’s Formula: ea+ib = ea ei b = ea cos(b) + iea sin(b).

• Given real and complex parts of a solution, the two parts can be treated as seperate
independent solutions and used in the linearization theorem to determine the general
solution.

• Stability: consider a linear two dimensional system with complex eigenvalues λ1 = a+ib
and λ2 = a − ib.

– If a is negative then solution spiral towards the origin (spiral sink).

– If a is positive then the solutions spiral away from the origin (spiral source).
– If a = 0 the solutions are periodic closed paths (neutral centers).

3.4 Repeated Eigenvalues

d
• Given the system, X = AX with one repeated eigenvalue, λ1 .
dt
• If V1 is an eigenvector, then X1 (t) = eλt V1 is a straight line solution.

• Another solution is of the form X2 (t) = eλt (tV1 + V2 ).

• Where V1 = (A − λI)V2 .

• X1 and X2 will be independent and the general solution is formed in the usual manner.

3.5 Zero as an Eigenvalue

• If zero is an eigenvector, nothing changes but the form of the general solution is now

X(t) = k1 V1 + k2 eλ2 t V2 .

7
4 Second Order Differential Equations
d2 y dy
• Form: 2
+ p(t) = q(t)y = f (t).
dt dt
• Homogeneous if f (t) = 0.

• given solutions y1 and y2 to the 2nd order differential equation, you must check the
Wronskian if both solutions are from real roots of the characteristic.

•  
y1 y2
W = det . (6)
y10 y20

• If W is equal to 0 anywhere on the interval of consideration, then y1 and y2 are not

linearly independent.

• General solution given y1 and y2 is found as usual by the linearization theorem.

• Characteristic polynomial of a 2nd order with constant coefficients: as2 + bs + c = 0.

• Solutions of the form y(t) = est .

√
b b2 − 4ac
• s=− +/− .
2a 2a
– if b2 − 4ac > 0, then two distinct real roots.
– if b2 − 4ac < 0, then complex roots.
– b2 − 4ac = 0, then repeated real roots.

4.1 Two real distinct Roots

• Two real roots, s1 and s2 .

• General solution = y(t) = k1 es1 t + k2 es−2t .

4.2 Complex Roots

• Complex Roots, s1 = p + iq and s2 = p − iq.

• General solution = y(t) = k1 ept cos(qt) + k2 ept sin(qt).

4.3 Repeated Roots

• Repeated Root, s1 .
b b
−t − t
• General solution = y(t) = k1 e 2a + k2 te a2 .

8
4.4 Nonhomogeneous with constant coefficents
• General solution = y(t) = yh + yp .

• Polynomial f (t).

– Look for particular solution of the form yp = Atn + Btn−1 + Ctn−2 + ... + Dt + E.

• Exponential f (t).

– Look for particular solution of the form yp = Aept .

• Sine or Cosine f (t).

– Look for particular solution of the form yp = Asin(at) + Bcos(at).

• Combination f (t).

– f (t) = Pn (t)eat , ⇒ yp = (Atn + Btn−1 + Ctn−2 + ... + Dt + E)eat .

– f (t) = Pn (t)sin(at) or Pn (t)cos(at), ⇒ yp = (A1tn + A2tn−1 + A3tn−2 + ... + A4t +
A5)cos(at) + (B1tn + B2tn−1 + B3tn−2 + ... + B4t + B5)sin(at).
– f (t) = eat sin(bt) or eat cos(bt), ⇒ yp = Aeat cos(bt) + Beat sin(bt).
– f (t) = Pn (t)eat sin(bt) or Pn (t)eat cos(bt), ⇒ yp = (A1tn + A2tn−1 + A3tn−2 + ... +
A4t + A5)eat cos(bt) + (B1tn + B2tn−1 + B3tn−2 + ... + B4t + B5)eat sin(bt).

• Superposition f (t).

– If f (t) is the sum of m terms of the forms previously described.

– yp = yp1 + yp2 + yp3 + ... + ypm .

9
5 LaPlace Transformations
R∞ RT
• Definition L{f (t)} = 0
e−st f (t)dt = limT →∞ 0
e−st f (t)dt.

• ONLY PROVIDED THAT THE INTEGRAL CONVERGES!!! MUST BE OF EX-

PONENTIAL ORDER!!!

• L{f (t)} = F (s).

1
• L{1} = .
s
1
• L{t} = .
s2
1
• L{eat } = .
s−a
ω
• L{sin(ωt)} = .
s2
+ ω2
s
• L{cos(ωt)} = 2 .
s + ω2
• Linear: L{αf (t) + βg(t)} = αF (s) + βG(s).

5.1 Inverse Laplace Transforms

• Linear: L−1 {αF (s) + βG(s)} = αf (t) + βg(t).

5.2 Transform of a derivative

• L{f 0 (t)} = sL(f (t) − f (0).

• L{f 00 (t)} = s2 L(f (t) − sf (0) − f 0 (0).

10