Solutions of the Sample Problems for the Third In-Class Exam

Math 246, Fall 2019, Professor David Levermore

(1) Compute the Laplace transform of f (t) = t e3t u(t − 2) from its definition.

Solution. The definition of the Laplace transform gives

Z T Z T
−st
L[f ](s) = lim 3t
e t e u(t − 2) dt = lim t e−(s−3)t dt .
T →∞ 0 T →∞ 2
This limit diverges to +∞ for s ≤ 3 because in that case for every T > 2 we have
Z T Z T
−(s−3)t T2
te dt ≥ t dt = − 2,
2 2 2
which clearly diverges to +∞ as T → ∞.
For s > 3 an integration by parts shows that
Z T T Z T −(s−3)t
−(s−3)t e−(s−3)t e
te dt = −t + dt
2 s−3 2 2 s−3
T
e−(s−3)t e−(s−3)t

= −t −
s−3 (s − 3)2 2
e−(s−3)T e−(s−3)T
  −(s−3)2
e−(s−3)2
 
e
= −T − + 2 + .
s−3 (s − 3)2 s−3 (s − 3)2
Hence, for s > 3 we have that
e−(s−3)T e−(s−3)T
  −(s−3)2
e−(s−3)2
 
e
L[f ](s) = lim −T − + 2 +
T →∞ s−3 (s − 3)2 s−3 (s − 3)2
e−(s−3)2 e−(s−3)2 e−(s−3)T e−(s−3)T
 
= +2 − lim T +
(s − 3)2 s−3 T →∞ s−3 (s − 3)2
e−(s−3)2 e−(s−3)2
= +2 .
(s − 3)2 s−3

(2) Consider the following MATLAB commands.

>> syms t y(t) s Y
>> f = heaviside(t)*tˆ2 + heaviside(t − 3)*(3*t − tˆ2);
>> diffeqn = diff(y, 2) − 6*diff(y, 1) + 10*y(t) == f;
>> eqntrans = laplace(diffeqn, t, s);
>> algeqn = subs(eqntrans, ...
[laplace(y(t), t, s), y(0), subs(diff(y(t), t), t, 0)], [Y, 2, 3]);
>> ytrans = simplify(solve(algeqn, Y));
>> y = ilaplace(ytrans, s, t)
(a) Give the initial-value problem for y(t) that is being solved.
(b) Find the Laplace transform Y (s) of the solution y(t).
1
2

DO NOT take the inverse Laplace transform of Y (s) to find y(t), just solve for Y (s)!
You may refer to the table on the last page.

Solution (a). The initial-value problem for y(t) that is being solved is
y 00 − 6y 0 + 10y = f (t) , y(0) = 2 , y 0 (0) = 3 ,
where the forcing f (t) can be expressed either as
(
t2 for 0 ≤ t < 3 ,
f (t) =
3t for 3 ≤ t ,
or in terms of the unit step function as f (t) = t2 + u(t − 3)(3t − t2 ).

Solution (b). The Laplace transform of the initial-value problem is

L[y 00 ](s) − 6L[y 0 ](s) + 10L[y](s) = L[f ](s) ,
where
L[y](s) = Y (s) ,
L[y 0 ](s) = s L[y](s) − y(0) = s Y (s) − 2 ,
L[y 00 ](s) = s L[y 0 ](s) − y 0 (0) = s2 Y (s) − 2s − 3 .
To compute L[f ](s), we first write f (t) as
f (t) = t2 + u(t − 3)(3t − t2 ) = t2 + u(t − 3)j(t − 3) ,
where by setting j(t − 3) = 3t − t2 we see by the shifty step method that
j(t) = 3(t + 3) − (t + 3)2 = 3t + 9 − t2 − 6t − 9 = −t2 − 3t .
Referring to the table on the last page, item 1 with a = 0 and n = 2 and with a = 0
and n = 1 shows that
2 1
L[t2 ](s) = 3 , L[t](s) = 2 ,
s s
2
whereby item 6 with c = 3 and j(t) = −t − 3t shows that
 
  −3s −3s
2  −3s 2 3
L u(t − 3)j(t − 3) (s) = e L[j](s) = −e L t + 3t (s) = −e + .
s3 s2
Therefore
 
2  2 −3s 2 3
L[f ](s) = L t + u(t − 3)j(t − 3) (s) = 3 − e + .
s s3 s2
The Laplace transform of the initial-value problem then becomes
 
2


2 −3s 2 3
s Y (s) − 2s − 3 − 6 sY (s) − 2 + 10Y (s) = 3 − e + ,
s s3 s2
which becomes
 
2 2 −3s 2 3
(s − 6s + 10)Y (s) − 2s + 9 = 3 − e + .
s s3 s2
Therefore Y (s) is given by
  
1 2 −3s 2 3
Y (s) = 2 2s − 9 + 3 − e + .
s − 6s + 10 s s3 s2
 3

(3) Find Y (s) = L[y](s) where y(t) solves the initial-value problem
y 00 + 4y 0 + 13y = f (t) , y(0) = 4 , y 0 (0) = 1 ,
where (
cos(t) for 0 ≤ t < 2π ,
f (t) =
t − 2π for t ≥ 2π .
DO NOT take the inverse Laplace transform of Y (s) to find y(t), just solve for Y (s)!
You may refer to the table on the last page.

Solution. The Laplace transform of the initial-value problem is

L[y 00 ](s) + 4L[y 0 ](s) + 13L[y](s) = L[f ](s) ,
where
L[y](s) = Y (s) ,
L[y 0 ](s) = s L[y](s) − y(0) = s Y (s) − 4 ,
L[y 00 ](s) = s L[y 0 ](s) − y 0 (0) = s2 Y (s) − 4s − 1 .
To compute L[f ](s), first write f as


f (t) = 1 − u(t − 2π) cos(t) + u(t − 2π)(t − 2π)


= cos(t) + u(t − 2π) t − 2π − cos(t)
= cos(t) + u(t − 2π)j(t − 2π) ,
where by setting j(t − 2π) = t − 2π − cos(t) we see by the shifty step method that
j(t) = (t + 2π) − 2π − cos(t + 2π) = t − cos(t) .
Here we have used the fact that cos(t) is 2π-periodic. Referring to the table on the
last page, item 6 with c = 2π shows that
L[f ](s) = L[cos(t)](s) + L[u(t − 2π)j(t − 2π)](s)
= L[cos(t)](s) + e−2πs L[j(t)](s)
= L[cos(t)](s) + e−2πs L[t − cos(t)](s) .
Then item 2 with a = 0 and b = 1, and item 1 with n = 1 and a = 1 imply that
 
s −2πs 1 s
L[f ](s) = 2 +e − .
s +1 s2 s2 + 1
The Laplace transform of the initial-value problem then becomes

s 1
s2 Y (s) − 4s − 1 + 4 sY (s) − 4 + 13Y (s) = 1 − e−2πs 2 + e−2πs 2 ,



s +1 s
which becomes

s 1
(s2 + 4s + 13)Y (s) − 4s − 1 − 16 = 1 − e−2πs 2 + e−2πs 2 .
s +1 s
Hence, Y (s) is given by
 
1 −2πs

s −2πs 1
Y (s) = 2 4s + 17 + 1 − e +e .
s + 4s + 13 s2 + 1 s2
4

(4) Find X(s) = L[x](s) where x(t) solves the initial-value problem
x00 + 4x = δ(t − 3) , x(0) = 5 , x0 (0) = 0 .
DO NOT take the inverse Laplace transform of X(s) to find x(t), just solve for X(s)!
You may refer to the table on the last page.

Solution. The Laplace transform of the initial-value problem is

L[x00 ](s) + 4L[x](s) = L[δ(t − 3)](s) ,
where
L[x](s) = X(s) ,
L[x0 ](s) = s L[x](s) − x(0) = s X(s) − 5 ,
L[x00 ](s) = s L[x0 ](s) − x0 (0) = s2 X(s) − 5s − 0 .
Referring to the table on the last page, item 7 with c = 3 and h(t) = 1 shows that
L[δ(t − 3)](s) = e−3s .
The Laplace transform of the initial-value problem then becomes
(s2 + 4)X(s) − 5s = e−3s .
Hence, X(s) is given by
5s + e−3s
X(s) = .
s2 + 4
Remark. You should be able to take the inverse Laplace transform to obtain
−3s
 
−1 −1 5s + e
x(t) = L [X](t) = L (t) = 5 cos(2t) + 21 u(t − 3) sin(2(t − 3)) .
s2 + 4

(5) Find the inverse Laplace transforms of the following functions.

2
(a) F (s) = ,
(s + 5)2
3s
(b) F (s) = ,
s2 −s−6
(s − 2)e−3s
(c) F (s) = 2 .
s − 4s + 5
You may refer to the table on the last page.

Solution (a). Referring to the table on the last page, item 1 with n = 1 and a = −5
gives
1
L[t e−5t ](s) = .
(s + 5)2
Therefore we conclude that
   
−1 2 −1 1
L 2
(t) = 2L 2
(t) = 2t e−5t .
(s + 5) (s + 5)
 5

Solution (b). Because the denominator factors as (s − 3)(s + 2), we have the partial
fraction identity
9 6
3s 3s 5 5
= = + .
s2 − s − 6 (s − 3)(s + 2) s−3 s+2
Referring to the table on the last page, item 1 with n = 0 and a = 3, and with n = 0
and a = −2 gives
1 1
L[e3t ](s) = , L[e−2t ](s) = .
s−3 s+2
Therefore we conclude that
   9 6 
−1 3s −1 5 5
L (t) = L + (t)
s2 − s − 6 s−3 s+2
   
9 −1 1 6 −1 1
= 5L (t) + 5 L (t)
s−3 s+2
= 59 e3t + 65 e−2t .

Solution (c). Complete the square in the denominator to get (s − 2)2 + 1. Referring
to the table on the last page, item 2 with a = 2 and b = 1 gives
s−2
L[e2t cos(t)](s) = .
(s − 2)2 + 1
Item 6 with c = 3 and j(t) = e2t cos(t) then gives
s−2
L[u(t − 3)e2(t−3) cos(t − 3)](s) = e−3s .
(s − 2)2 + 1
Therefore we conclude that
 
−1 −3s s−2
L e (t) = u(t − 3)e2(t−3) cos(t − 3) .
s2 − 4s + 5

(6) For each of the following differential operators compute its Green function g(t) and
its natural fundamental set for t = 0.
(a) L = D4 + 8D2 − 9 ,
(b) L = (D − 2)3 .
You may refer to the table on the last page.

Solution (a). The characteristic polynomial of L = D4 +8D2 −9 is p(s) = s4 +8s2 −9.

Therefore its Green function g(t) is given by
   
−1 1 −1 1
g(t) = L (t) = L (t) .
p(s) s4 + 8s2 − 9
Because p(s) factors as p(s) = (s2 − 1)(s2 + 9) we have the partial fraction identity
1 1
1 1 10
− 10
= 2 = 2 + .
s4 + 8s2 − 9 (s − 1)(s2 + 9) s − 1 s2 + 9
6

Because s2 − 1 factors as s2 − 1 = (s − 1)(s + 1) we have the partial fraction identity

1
1 1 2
− 12
= = + .
s2 − 1 (s − 1)(s + 1) s−1 s+1
By combining the above partial fraction identities we obtain
1 1 1 1 1 1 1
4 2
= 20 − 20 − 10 2
.
s + 8s − 9 s−1 s+1 s +9
Referring to the table on the last page, item 1 with a = 1 and n = 0 and with a = −1
and n = 0 gives
   
−1 1 −1 1
L t
(t) = e , L (t) = e−t ,
s−1 s+1
while item 3 with a = 0 and b = 3 gives
 
−1 3
L (t) = sin(3t) .
s2 + 9
Therefore the Green function g(t) is given by
 
−1 1
g(t) = L (t)
s4 + 8s2 − 9
     
1 −1 1 1 −1 1 1 −1 3
= 20 L (t) − 20 L (t) − 30
L (t)
s−1 s+1 s2 + 9
1 t 1 −t 1
= 20 e − 20 e − 30 sin(3t) .
Then because we see the characteristic polynomial as
p(s) = s4 + 0s3 + 8s2 + 0s − 9 ,
the natrual fundamental set for t = 0 is found by
1 t 1 −t 1
N3 (t) = g(t) = 20
e − 20
e − 30 sin(3t) ,
N2 (t) = N30 (t) + 0g(t) = 201 t 1 −t
e + 20 e − 10 1
cos(3t) ,
N1 (t) = N20 (t) + 8g(t)
1 t 1 −t 3 8 t 8 −t 8
= 20
e − 20 e + 10 sin(3t) + 20 e − 20 e − 30 sin(3t) ,
9 t 9 −t 1
= 20
e − 20 e + 30 sin(3t) ,
N0 (t) = N10 (t) + 0g(t) = 20
9 t 9 −t
e + 20 1
e + 10 cos(3t) .
Remark. The calculation of the natural fundamental set is a bit simpler if the Green
function is expressed in terms of hyperbolic functions. It becomes
1 1
N3 (t) = g(t) = 10
sinh(t) − 30
sin(3t) ,
N2 (t) = N30 (t) + 0g(t) = 101 1
cosh(t) − 10 cos(3t) ,
N1 (t) = N20 (t) + 8g(t)
1 3 8 8
= 10
sinh(t) + 10 sin(3t) + 10 sinh(t) − 30 sin(3t) ,
9 1
= 10
sinh(t) + 30 sin(3t) ,
0 9 1
N0 (t) = N1 (t) + 0g(t) = 10 cosh(t) + 10 cos(3t) .
 7

Solution (b). The characteristic polynomial of L = (D − 2)3 is p(s) = (s − 2)3 .

Therefore its Green function g(t) is given by
   
−1 1 −1 1
g(t) = L (t) = L (t) .
p(s) (s − 2)3
Referring to the table on the last page, item 1 with a = 2 and n = 2 gives
 
1 −1 2
g(t) = 2 L (t) = 21 t2 e2t .
(s − 2)3
Then because by the binomial expansion we see the characteristic polynomial as
p(s) = (s − 2)3 = s3 + 3(−2)s2 + 3(−2)2 s + (−2)3
= s3 − 6s2 + 12s − 8 ,
the natrual fundamental set for t = 0 is found by
N2 (t) = g(t) = 21 t2 e2t ,
N1 (t) = N20 (t) − 6g(t)
= t e2t + t2 e2t − 26 t2 e2t = t e2t − 2t2 e2t ,

N0 (t) = N10 (t) + 12g(t)

= e2t − 2t e2t − 4t2 e2t + 12 2 2t
= e2t − 2t e2t + 2t2 e2t .


2
te

(7) Recast the equation u000 + t2 u0 − 3u = sinh(2t) as a first-order system of ordinary

differential equations.
Solution. Because the equation is third order, the first-order system must have
dimension three. The simplest such first-order system is
       
x1 x2 x1 u
d     , where x2  =  u0  .
x2 = x3
dt x sinh(2t) + 3x − t2 x x u00
3 1 2 3

(8) Two interconnected tanks are filled with brine (salt water). At t = 0 the first tank
contains 45 liters and the second contains 30 liters. Brine with a salt concentration
of 5 grams per liter flows into the first tank at 6 liters per hour. Well-stirred brine
flows from the first tank into the second at 8 liters per hour, from the second into
the first at 7 liters per hour, from the first into a drain at 4 liter per hour, and from
the second into a drain at 3 liters per hour. At t = 0 there are 27 grams of salt in
the first tank and 18 grams in the second.
(a) Give an initial-value problem that governs the amount of salt in each tank as a
function of time.
(b) Give the interval of definition for the solution of this initial-value problem.
8

Solution (a). Let V1 (t) and V2 (t) be the volumes (lit) of brine in the first and second
tank at time t hours. Let S1 (t) and S2 (t) be the mass (gr) of salt in the first and
second tank at time t hours. Because mixtures are assumed to be well-stirred, the
salt concentration of the brine in the tanks at time t are C1 (t) = S1 (t)/V1 (t) and
C2 (t) = S2 (t)/V2 (t) respectively. In particular, these are the concentrations of the
brine that flows out of these tanks. We have the following picture.

5 gr/lit C1 (t) gr/lit

→ → →
6 lit/hr 8 lit/hr
V1 (t) lit V2 (t) lit
S1 (t) gr S2 (t) gr
C1 (t) gr/lit C2 (t) gr/lit C2 (t) gr/lit
← ← ← →
4 lit/hr 7 lit/hr 3 lit/hr
V1 (0) = 45 lit V2 (0) = 30 lit
S1 (0) = 27 gr S2 (0) = 18 gr
We are asked to write down an initial-value problem that governs S1 (t) and S2 (t).
The rates work out so there will be V1 (t) = 45 + t liters of brine in the first tank
and V2 (t) = 30 − 2t liters in the second. Then S1 (t) and S2 (t) are governed by the
initial-value problem
dS1 S2 S1 S1
=5·6+ 7− 8− 4, S1 (0) = 27 ,
dt 30 − 2t 45 + t 45 + t
dS2 S1 S2 S2
= 8− 7− 3, S2 (0) = 18 .
dt 45 + t 30 − 2t 30 − 2t
You could leave the answer in the above form. However, it can be simplified to
dS1 7 12
= 30 + S2 − S1 , S1 (0) = 27 ,
dt 30 − 2t 45 + t
dS2 8 5
= S1 − S2 , S2 (0) = 18 .
dt 45 + t 15 − t

Solution (b). This first-order system of differential equations is linear.

 Its coefficients are undefined either at t = −45 or at t = 15 and are continuous
elsewhere.
 Its forcing is constant, so is continuous everywhere.
 Its initial time is t = 0.
Therefore the natural interval of definition for the solution of this initial-value prob-
lem is (−45, 15) because:
• the initial time t = 0 is in (−45, 15);
• all the coefficients and the forcing are continuous over (−45, 15);
• two coefficients are undefined at t = −45;
• two coefficients are undefined at t = 15.
However, it could also be argued that the interval of definition for the solution of this
initial-value problem is [0, 15) because the word problem starts at t = 0.
 9

(9) Consider the matrices

   
−i2 1 + i 7 6
A= , B= .
2 + i −4 8 7
Compute the matrices
(a) AT ,
(b) A ,
(c) AH ,
(d) 5A − B ,
(e) AB ,
(f) B−1 .

Solution (a). The transpose of A is

 
T −i2 2 + i
A = .
1 + i −4

Solution (b). The conjugate of A is

 
i2 1 − i
A= .
2 − i −4

Solution (c). The Hermitian transpose of A is

 
H i2 2 − i
A = .
1 − i −4

Solution (d). The difference of 5A and B is given by

     
−i10 5 + i5 7 6 −7 − i10 −1 + i5
5A − B = − = .
10 + i5 −20 8 7 2 + i5 −27

Solution (e). The product of A and B is given by

  
−i2 1 + i 7 6
AB =
2 + i −4 8 7
 
−i2·7 + (1 + i)·8 −i2·6 + (1 + i)·7
=
(2 + i)·7 − 4·8 (2 + i)·6 − 4·7
 
8 − i6 7 − i5
= .
−18 + i7 −16 + i6

Solution (f ). Observe that it is clear that B has an inverse because

 
7 6
det(B) = det = 7 · 7 − 6 · 8 = 49 − 48 = 1 .
8 7
Then the inverse of B is given by
   
−1 1 7 −6 7 −6
B = = .
det(B) −8 7 −8 7
10
4   2
t +3 t
(10) Consider the vector-valued functions x1 (t) = 2 , x2 (t) = .
2t 3
(a) Compute the Wronskian Wr[x1 , x2 ](t).
dx
(b) Find A(t) such that x1 , x2 is a fundamental set of solutions to = A(t)x
dt
wherever Wr[x1 , x2 ](t) 6= 0.
(c) Give a fundamental matrix Ψ(t) for the system found in part (b).
(d) For the system found in part (b), solve the initial-value problem
 
dx 1
= A(t)x , x(1) = .
dt 0
(e) For the A(t) found in part (b), give the Green matrix for the system
x0 = A(t)x + f (t) .

Solution (a).
4 
t + 3 t2
Wr[x1 , x2 ](t) = det = 3t4 + 9 − 2t4 = t4 + 9 .
2t2 3

4 
t + 3 t2 dΨ(t)
Solution (b). Let Ψ(t) = . Because = A(t)Ψ(t), we have
2t2 3 dt
 3  4 −1
dΨ(t) −1 4t 2t t +3 t2
A(t) = Ψ(t) =
dt 4t 0 2t2 3
 3    3 
1 4t 2t 3 −t2 1 8t 6t − 2t5
= 4 = .
t + 9 4t 0 −2t2 t4 + 3 t4 + 9 12t −4t3

Solution (c). Because x1 (t), x2 (t) is a fundamental set of solutions to the system
found in part (b), a fundamental matrix for the system found in part (b) is simply
given by
4 

t + 3 t2
Ψ(t) = x1 (t) x2 (t) = .
2t2 3

Solution (d). Because a fundamental matrix Ψ(t) for the system found in part (b)
was given in part (c), the solution of the initial-value problem is
4  −1  
−1 t + 3 t2 4 1 1
x(t) = Ψ(t)Ψ(1) x(1) = 2
2t 3 2 3 0
4 2
   
t +3 t 1 3 −1 1
= 2
2t 3 10 −2 4 0
4     
1 t + 3 t2 3 1 3t4 + 9 − 2t2
= = .
10 2t2 3 −2 10 6t2 − 6
 11

Alternative Solution (d). Because x1 (t), x2 (t) is a fundamental set of solutions

to the system found in part (b), a general solution is given by
4   2
t +3 t
x(t) = c1 x1 (t) + c2 x2 (t) = c1 2 + c2 .
2t 3
The initial condition then implies that
       
4 1 4c1 + c2 1
x(1) = c1 + c2 = = ,
2 3 2c1 + 3c2 0
3
from which we see that c1 = 10 and c2 = − 15 . The solution of the initial-value problem
is thereby
4   2  3 4 1 2 9

3 t +3 1 t t − t +
x(t) = 10 −5 = 10 3 2 5 3 10 .
2t2 3 5
t −5

Solution (e). Because a fundamental matrix Ψ(t) for the system found in part (b)
was given in part (c), the Green matrix for the nonhomogeneous system is
4  4 −1
−1 t + 3 t2 s + 3 s2
G(t, s) = Ψ(t)Ψ(s) =
2t2 3 2s2 3
4 2
  2

t +3 t 1 3 −s
= 2
2t 3 s + 9 −2s s4 + 3
4 2
 4 
1 3t + 9 − 2t2 s2 t2 (s4 + 3) − (t4 + 3)s2
= 4 .
s +9 6t2 − 6s2 3s4 + 9 − 2t2 s2

etA forthe following matrices.

(11) Compute 
1 4
(a) A =
1 1
 
6 4
(b) A =
−1 2

Solution (a) by Two-by-Two Formula. Because

 
1 4
A= ,
1 1
the characteristic polynomial of A is given by
p(z) = z 2 − tr(A)z + det(A) = z 2 − 2z − 3 = (z − 1)2 − 22 .
This is a difference of squares with µ = 1 and ν = 2. Hence,
 
tA t sinh(2t)
e = e cosh(2t)I + (A − I)
2
    
t 1 0 sinh(2t) 0 4
= e cosh(2t) +
0 1 2 1 0
 
cosh(2t) 2 sinh(2t)
= et 1 .
2
sinh(2t) cosh(2t)
12

Solution (a) by the Natural Fundamental Set Method. Because

 
1 4
A= ,
1 1
the characteristic polynomial of A is
p(z) = z 2 − tr(A)z + det(A) = z 2 − 2z − 3 = (z + 1)(z − 3) .
Below we show that the natural fundamental set of solutions for t = 0 associated
with p(D) = D2 − 2D + 3 is
e3t + 3e−t e3t − e−t
N0 (t) = , N1 (t) = .
4 4
Then
e3t + 3e−t 1 e3t − e−t 1 4
   
tA 0
e = N0 (t)I + N1 (t)A = +
4 0 1 4 1 1
1 2e + 2e−t 4e3t − 4e−t
 3t 
= .
4 e3t − e−t 2e3t + 2e−t

From the Green Function. By the partial fraction identity

1
1 1 4
− 14
= = + ,
s2 − 2s + 3 (s − 3)(s + 1) s−3 s+1
the Green function associated with p(D) = D2 − 2D + 3 is
   
−1 1 −1 1
g(t) = L (t) = L (t)
p(s) s2 − 2s + 3
   
1 −1 1 1 −1 1
= 4L (t) − 4 L (t) = 14 e3t − 14 e−t .
s−3 s+1
Then, because the characteristic polynomial is p(s) = s2 − 2s + 3, the natural funda-
mental set is
N1 (t) = g(t) = 41 e3t − 41 e−t ,
N0 (t) = N10 (t) − 2g(t) = 34 e3t + 14 e−t − 42 e3t − 42 e−t = 14 e3t + 43 e−t .

From the General Initial-Value Problem. The general initial-value problem

associated with p(D) = D2 − 2D + 3 is
y 00 − 2y 0 − 3y = 0 , y(0) = y0 , y 0 (0) = y1 .
This has the general solution y(t) = c1 e3t + c2 e−t . Because y 0 (t) = 3c1 e3t − c2 e−t , the
general initial conditions yield
y0 = y(0) = c1 + c2 , y1 = y 0 (0) = 3c1 − c2 .
This system can be solved to obtain
y0 + y1 3y0 − y1
c1 = , c2 = .
4 4
The solution of the general initial-value problem is thereby
y0 + y1 3t 3y0 − y1 −t e3t + 3e−t e3t − e−t
y(t) = e + e = y0 + y1 .
4 4 4 4
 13

Therefore the associated natural fundamental set of solutions is

e3t + 3e−t e3t − e−t
N0 (t) = , N1 (t) = .
4 4

Solution (a) by Eigen Methods. Because

 
1 4
A= ,
1 1
the characteristic polynomial of A is
p(z) = z 2 − tr(A)z + det(A) = z 2 − 2z − 3 = (z + 1)(z − 3) .
The eigenvalues of A are the roots of this polynomial, which are −1 and 3. Because
   
2 4 −2 4
A+I= , A − 3I = ,
1 2 1 −2
we can read off that A has the eigenpairs
     
2 2
−1 , , 3, .
−1 1
Set    
2 2 −1 0
V= , D= .
−1 1 0 3
Because det(V) = 4, we see that
   −t  −1
tA tD −1 2 2 e 0 2 2
e = Ve V =
−1 1 0 e3t −1 1
  −t   −t
−2e−t
     
2 2 e 0 1 1 −2 1 2 2 e
= =
−1 1 0 e3t 4 1 2 4 −1 1 e3t 2e3t
1 2e−t + 2e3t 4e3t − 4e−t
 
= .
4 e3t − e−t 2e−t + 2e3t

Solution (b) by Two-by-Two Formula. Because

 
6 4
A= ,
−1 2
the characteristic polynomial of A is given by
p(z) = z 2 − tr(A)z + det(A) = z 2 − 8z + 16 = (z − 4)2 .
This is a perfect square with µ = 4. Hence,
     
tA 4t

4t 1 0 2 4
e = e I + t A − 4I = e +t
0 1 −1 −2
 
4t 1 + 2t 4t
=e .
−t 1 − 2t

Solution (b) by the Natural Fundamental Set Method. Because

 
6 4
A= ,
−1 2
14

the characteristic polynomial of A is

p(z) = z 2 − tr(A)z + det(A) = z 2 − 8z + 16 = (z − 4)4 .
Below we show that the natural fundamental set of solutions for t = 0 associated
with p(D) = D2 − 8D + 16 is
N0 (t) = (1 − 4t)e4t , N1 (t) = t e4t .
Then
   
tA 1 04t 4t 6 4
e = N0 (t)I + N1 (t)A = (1 − 4t)e + te
0 1 −1 2
 
1 + 2t 4t
= e4t .
−t 1 − 2t

From the Green Function. Because p(s) = s2 − 8s + 16 = (s − 4)2 , the Green

function associated with p(D) = D2 − 8D + 16 is
   
−1 1 −1 1
g(t) = L (t) = L (t)
p(s) s2 − 8s + 16
 
−1 1
=L (t) = t e4t .
(s − 4)2
Then, because the characteristic polynomial is p(s) = s2 − 8s + 16, the natural
fundamental set is
N1 (t) = g(t) = t e4t ,
N0 (t) = N10 (t) − 8g(t) = e4t + 4t e4t − 8t e4t = e4t − 4t e4t .

From the General Initial-Value Problem. The general initial-value problem

associated with p(D) = D2 − 8D + 16 is
y 00 − 8y 0 + 16y = 0 , y(0) = y0 , y 0 (0) = y1 .
This has the general solution y(t) = c1 e4t + c2 te4t . Because
y 0 (t) = 4c1 e3t + 4c2 te−t + c2 e4t ,
the general initial conditions yield
y0 = y(0) = c1 , y1 = y 0 (0) = 4c1 + c2 .
This system can be solved to obtain c1 = y0 and c2 = y1 − 4y0 . The solution of the
general initial-value problem is thereby
y(t) = y0 e4t + (y1 − 4y0 )t e4t = (1 − 4t)e4t y0 + t e4t y1 .
Therefore the associated natural fundamental set of solutions is
N0 (t) = (1 − 4t)e4t , N1 (t) = t e4t .

(12) Give the Green matrix for the system x0 = Ax + f (t) when
 
1 4
(a) A =
1 1
 15
 
6 4
(b) A =
−1 2

Solution (a). By the solution to part (a) of the previous problem

1 2e3t + 2e−t 4e3t − 4e−t
 
tA
e = .
4 e3t − e−t 2e3t + 2e−t
Therefore the Green matrix G(t, s) is given by
1 2e3(t−s) + 2e−(t−s) 4e3(t−s) − 4e−(t−s)
 
tA −sA (t−s)A
G(t, s) = e e =e = .
4 e3(t−s) − e−(t−s) 2e3(t−s) + 2e−(t−s)

Solution (b). By the solution to part (b) of the previous problem

 
tA 4t 1 + 2t 4t
e =e .
−t 1 − 2t

Therefore the Green matrix G(t, s) is given by

 
tA −sA (t−s)A 4(t−s) 1 + 2(t − s) 4(t − s)
G(t, s) = e e =e =e .
−(t − s) 1 − 2(t − s)

(13) Consider the matrix

 
−1 −2 1
A= 4 0 −2 .
−2 0 1

Compute etA given that the characteristic polynomial of A is p(z) = z 3 + 9z and that
the natural fundamental set of solutions associated with t = 0 for D3 + 9D is
N1 (t) = 13 sin(3t) , N2 (t) = 19 1 − cos(3t) .


N0 (t) = 1 ,

Solution. The natural fundamental set method says that

etA = N0 (t)I + N1 (t)A + N2 (t)A2 .
Because N0 (t) = 1, N1 (t) = 31 sin(3t), N2 (t) = 91 1 − cos(3t) , and

  
−1 −2 1 −1 −2 1
A2 =  4 0 −2  4 0 −2
−2 0 1 −2 0 1
   
1−8−2 2 −1 + 4 + 1 −9 2 4
= −4 + 0 + 4 −8 4 − 0 − 2  =  0 −8 2  ,
2+0−2 4 −2 − 0 + 1 0 4 −1
16

we see that
     
1 0 0 −1 −2 1 −9 2 4
etA = 1 0 1 0 + 13 sin(3t)  4 0 −2 + 19 1 − cos(3t)  0 −8 2 


0 0 1 −2 0 1 0 4 −1
 
cos(3t) − 31 sin(3t) − 23 sin(3t) + 32 − 23 cos(3t) 31 sin(3t) + 49 − 94 cos(3t)
4 1
= 3
sin(3t) 9
+ 89 cos(3t) − 23 sin(3t) + 92 − 29 cos(3t) .
− 23 sin(3t) 4
9
− 49 cos(3t) 8
9
+ 91 cos(3t) + 13 sin(3t)

(14) Solve each of the following initial-value problems.

        
d x 2 2 x x(0) 1
(a) = , = .
dt y 5 −1 y y(0) −1
        
d x 1 1 x x(0) 1
(b) = , = .
dt y −4 1 y y(0) 1
 
2 2
Solution (a). The characteristic polynomial of A = is given by
5 −1

p(z) = z 2 − tr(A)z + det(A) = z 2 − z − 12 = (z − 12 )2 − ( 27 )2 .

This is a difference of squares with µ = 12 and ν = 72 . Hence,
sinh 72 t


1
tA t 7 1



e = e 2 cosh 2 t I + 7 A − 2I
2
" #
7


3
 
1 1 0 sinh t 2
= e 2 t cosh 72 t 2


+ 2
0 1 7
2
5 − 32
7

3 7

4 7


1
t cosh 2 t + 7 sinh 2
t 7
sinh 2
t
=e 2
10 .
sinh 72 t cosh 72 t − 37 sinh 7




7 2
t
Therefore the solution of the initial-value problem is
     
x(t) tA x(0) tA 1
=e =e
y(t) y(0) −1
7

3 7

4 7

 
1
t cosh 2 t + 7 sinh 2
t 7
sinh 2
t 1
=e 2
10
sinh 72 t cosh 72 t − 73 sinh 7




7 2
t −1
7

1 7


1
t cosh 2
t − 7
sinh 2
t
= e2
.
− cosh 72 t + 13 7


7
sinh 2
t

 
1 1
Solution (b). The characteristic polynomial of A = is given by
−4 1

p(z) = z 2 − tr(A)z + det(A) = z 2 − 2z + 5 = (z − 1)2 + 22 .

 17

This is a sum of squares with µ = 1 and ν = 2. Hence,

 
tA t sin(2t)
e = e cos(2t)I + (A − I)
2
    
t 1 0 sin(2t) 0 1
= e cos(2t) +
0 1 2 −4 0
1
 
cos(2t) sin(2t)
= et 2 .
−2 sin(2t) cos(2t)
Therefore the solution of the initial-value problem is
     
x(t) tA x(0) tA 1
=e =e
y(t) y(0) 1
1
cos(2t) + 12 sin(2t)
    
t cos(2t) 2
sin(2t) 1 t
=e =e .
−2 sin(2t) cos(2t) 1 −2 sin(2t) + cos(2t)

Remark. We could have used other methods to compute etA in each part of the
above problem. Alternatively, we could have constructed a fundamental matrix Ψ(t)
and then determined c so that Ψ(t)c satisfies the initial conditions.

(15) Find a general solution for each of the following systems.

    
d x 3 −4 x
(a) =
dt y 1 −1 y
    
d x 2 −5 x
(b) =
dt y 4 −2 y
    
d x 5 4 x
(c) =
dt y −5 1 y

Solution (a). We must find a general solution for the system

    
d x 3 −4 x
= .
dt y 1 −1 y
 
3 −4
The characteristic polynomial of A = is given by
1 −1
p(z) = z 2 − tr(A)z + det(A) = z 2 − 2z + 1 = (z − 1)2 .
This is a perfect square with µ = 2. Hence,
   
tA t
  t 1 0 2 −4
e = e I + t (A − I) = e +t
0 1 1 −2
 
t 1 + 2t −4t
=e .
t 1 − 2t
18

Therefore a general solution is

      
x(t) tA c1 t 1 + 2t −4t c1
=e =e
y(t) c2 t 1 − 2t c2
   
1 + 2t −4t
= c1 e t + c2 et .
t 1 − 2t

Solution (b) by Two-by-Two Formula. We must find a general solution for the
system
    
d x 2 −5 x
= .
dt y 4 −2 y
 
2 −5
The characteristic polynomial of A = is given by
4 −2
p(z) = z 2 − tr(A)z + det(A) = z 2 + 42 .
This is a sum of squares with µ = 0 and ν = 4. Hence,
      
tA sin(4t) 1 0 sin(4t) 2 −5
e = cos(4t)I + A = cos(4t) +
4 0 1 4 4 −2
cos(4t) + 12 sin(4t) − 54 sin(4t)
 
= .
sin(4t) cos(4t) − 21 sin(4t)
Therefore a general solution is
cos(4t) + 21 sin(4t) − 54 sin(4t)
      
x(t) tA c1 c1
=e = 1
y(t) c2 sin(4t) cos(4t) − 2 sin(4t) c2
cos(4t) + 12 sin(4t) − 45 sin(4t)
   
= c1 + c2 .
sin(4t) cos(4t) − 12 sin(4t)

Solution (b) by Eigen Methods. We must find a general solution for the system
    
d x 2 −5 x
= .
dt y 4 −2 y
 
2 −5
The characteristic polynomial of A = is
4 −2
p(z) = z 2 − tr(A)z + det(A) = z 2 + 16 .
The eigenvalues of A are the roots of this polynomial, which are ±i4. Because
   
2 − i4 −5 2 + i4 −5
A − i4I = , A + i4I = ,
4 −2 − i4 4 −2 + i4
we can read off that A has the eigenpairs
     
1 + i2 1 − i2
i4 , , −i4 , .
2 2
 19

Therefore the system has the complex-valued solution

   
i4t 1 + i2

1 + i2
e = cos(4t) + i sin(4t)
2 2
 
cos(4t) − 2 sin(4t) + i2 cos(4t) + i sin(4t)
= .
2 cos(4t) + i2 sin(4t)
By taking real and imaginary parts, we obtain the two real solutions
    
i4t 1 + i2 cos(4t) − 2 sin(4t)
x1 (t) = Re e = ,
2 2 cos(4t)
    
i4t 1 + i2 2 cos(4t) + sin(4t)
x2 (t) = Im e = .
2 2 sin(4t)
Therefore a general solution is
     
x(t) cos(4t) − 2 sin(4t) 2 cos(4t) + sin(4t)
= c1 + c2 .
y(t) 2 cos(4t) 2 sin(4t)

Solution (c) by Two-by-Two Formula. We must find a general solution for the
system
    
d x 5 4 x
= .
dt y −5 1 y
 
5 4
The characteristic polynomial of A = is given by
−5 1
p(z) = z 2 − tr(A)z + det(A) = z 2 − 6z + 25 = (z − 3)2 + 42 .
This is a sum of squares with µ = 3 and ν = 4. Hence,
 
tA 3t sin(4t)
e = e cos(4t)I + (A − 3I)
4
    
3t 1 0 sin(4t) 2 4
= e cos(4t) +
0 1 4 −5 −2
1
 
3t cos(4t) + 2 sin(4t) sin(4t)
=e .
− 45 sin(4t) cos(4t) − 12 sin(4t)
Therefore a general solution is
1
      
x(t) tA c1 3t cos(4t) + 2 sin(4t) sin(4t) c1
=e =e 5 1
y(t) c2 − 4 sin(4t) cos(4t) − 2 sin(4t) c2
cos(4t) + 21 sin(4t)
   
sin(4t)
= c1 e3t + c2 e3t .
− 54 sin(4t) cos(4t) − 12 sin(4t)

Solution (c) by Eigen Methods. We must find a general solution for the system
    
d x 5 4 x
= .
dt y −5 1 y
20
 
5 4
The characteristic polynomial of A = is
−5 1
p(z) = z 2 − tr(A)z + det(A) = z 2 − 6z + 25 = (z − 3)2 + 16 .
The eigenvalues of A are the roots of this polynomial, which are 3 ± i4. Because
   
2 − i4 4 2 + i4 4
A − (3 + i4)I = , A − (3 − i4)I = ,
−5 −2 − i4 −5 −2 + i4
we can read off that A has the eigenpairs
     
−2 −2
3 + i4 , , 3 − i4 , .
1 − i2 1 + i2
Therefore the system has the complex-valued solution
   
(3+i4)t −2 3t

−2
e = e cos(4t) + i sin(4t)
1 − i2 1 − i2
 
3t −2 cos(4t) − i2 sin(4t)
=e .
cos(4t) + 2 sin(4t) + i sin(4t) − i2 cos(4t)
By taking real and imaginary parts, we obtain the two real solutions
    
(3+i4)t −2 3t −2 cos(4t)
x1 (t) = Re e =e ,
1 − i2 cos(4t) + 2 sin(4t)
    
(3+i4)t −2 3t −2 sin(4t)
x2 (t) = Im e =e .
1 + i2 sin(4t) − 2 cos(4t)
Therefore a general solution is
     
x(t) 3t −2 cos(4t) 3t −2 sin(4t)
= c1 e + c2 e .
y(t) cos(4t) + 2 sin(4t) sin(4t) − 2 cos(4t)

(16) Given that 1 is an eigenvalue of the matrix

 
2 −1 1
A = 1 1 −1 ,
0 −1 3
find all the eigenvectors of A associated with 1.

Solution. The eigenvectors of A associated with 1 are all nonzero vectors v that
satisfy Av = v. Equivalently, they are all nonzero vectors v that satisfy (A−I)v = 0,
which is   
1 −1 1 v1
1 0 −1 v2  = 0 .
0 −1 2 v3
The entries of v thereby satisfy the homogeneous linear algebraic system
v1 − v2 + v3 = 0 ,
v1 − v3 = 0 ,
− v2 + 2v3 = 0 .
 21

We may solve this system either by elimination or by row reduction. By either

method we find that its general solution is
v1 = α , v2 = 2α , v3 = α , for any constant α .
The eigenvectors of A associated with 1 thereby have the form
 
1
α 2 for any nonzero constant α .
1

(17) Consider the matrix

 
3 3
A= .
4 −1
(a) Find all the eigenvalues of A.
(b) For each eigenvalue of A find all of its eigenvectors.
(c) Diagonalize A.
(d) Compute etA .
(e) Compute (sI − A)−1 for every s where it is defined.

Solution (a). The characteristic polynomial of A is given by

p(z) = z 2 − tr(A)z + det(A) = z 2 − 2z − 15 = (z − 1)2 − 16 .
The eigenvalues of A are the roots of this polynomial, which are 1 ± 4, or simply −3
and 5.

Solution (b) by the Cayley-Hamilton Method. We have

   
6 3 −2 3
A + 3I = , A − 5I = .
4 2 4 −6
Every nonzero column of A − 5I has the form
 
1
α1 for some α1 6= 0 .
−2
These are all the eigenvectors associated with −3. Similarly, every nonzero column
of A + 3I has the form
 
3
α2 for some α2 6= 0 .
2
These are all the eigenvectors associated with 5.

Solution (c). If we use the eigenpairs

     
1 3
−3 , , 5, ,
−2 2
then set    
1 3 −3 0
V= , D= .
−2 2 0 5
22

Because det(V) = 1 · 2 − (−2) · 3 = 2 + 6 = 8, we see that

 
−1 1 2 −3
V = .
8 2 1
We conclude that A has the diagonalization
    
−1 1 3 −3 0 1 2 −3
A = VDV = .
−2 2 0 5 8 2 1
You do not have to multiply these matrices out. Had we started with different
eigenpairs, the steps would be the same as above but we would obtain a different
diagonalization.

Solution (d). Because A = VDV−1 by part (c), we have

   −3t   
tA tD −1 1 3 e 0 1 2 −3
e = Ve V =
−2 2 0 e5t 8 2 1
  −3t
−3e−3t
 
1 1 3 2e
=
8 −2 2 2e5t e5t
1 2e + 6e5t −3e−3t + 3e5t
 −3t 
= .
8 −4e−3t + 4e5t 6e−3t + 2e5t

Solution (e). Because A = VDV−1 by part (c), we have

1
 
 
1 3 s + 3 0  1 2 −3
−1 −1 −1
(sI − A) = V(sI − D) V = 1 8 2 1
−2 2 
0
s−5
2 −3
 
 
1 1 3 s + 3 s + 3
=  2 1 
8 −2 2
s−5 s−5
2 6 −3 3
 
1  + 3 + s − 5 s + 3 + s − 5
=  s−4 4 6 2 .
8 + +
s+3 s−5 s+3 s−5
This is defined for every s except at s = −3 and s = 5.

(18) What answer will be produced by the following Matlab command?

>> A = [1 4; 3 2]; [vect, val] = eig(sym(A))
You do not have to give the answer in Matlab format.
 
1 4
Solution. The Matlab command will produce the eigenpairs of A = . The
3 2
characteristic polynomial of A is
p(z) = z 2 − tr(A)z + det(A) = z 2 − 3z − 10 = (z − 5)(z + 2) ,
 23

so its eigenvalues are 5 and −2. Because

   
−4 4 3 4
A − 5I = , A + 2I = ,
3 −3 3 4
we can read off that the eigenpairs are
     
1 −4
5, , −2 , .
1 3

(19) A 3 × 3 matrix A has the eigenpairs

       

1 −1 1
−3 , 1 , 2 ,  1  , 5 , −1 .
0 1 2
(a) Give an invertible matrix V and a diagonal matrix D such that etA = VetD V−1 .
(You do not have to compute either V−1 or etA !)
(b) Give a fundamental matrix for the system x0 = Ax.
Solution (a). One choice for V and D is
   
1 −1 1 −3 0 0
V = 1 1 −1 , D =  0 2 0 .
0 1 2 0 0 5

Solution (b). Use the given eigenpairs to construct the special solutions
     
1 −1 1
x1 (t) = e−3t 1 , x2 (t) = e2t  1  , x3 (t) = e5t −1 ,
0 1 2
Then a fundamental matrix for the system is
 −3t 
e −e2t e5t
Ψ(t) = x1 (t) x2 (t) x3 (t) = e−3t e2t −e5t  .


0 e2t 2e5t

Alternative Solution (b). Given the V and D from part (a), a fundamental matrix
for the system is
   −3t   −3t 
1 −1 1 e 0 0 e −e2t e5t
Ψ(t) = VetD = 1 1 −1  0 e2t 0  = e−3t e2t −e5t  .
0 1 2 0 0 e5t 0 e2t 2e5t
24

Table of Laplace Transforms

n!
L[tn eat ](s) = for s > a .
(s − a)n+1
s−a
L[eat cos(bt)](s) = for s > a .
(s − a)2 + b2
b
L[eat sin(bt)](s) = for s > a .
(s − a)2 + b2

L[j 0 (t)](s) = sJ(s) − j(0) where J(s) = L[j(t)](s) .

L[tn j(t)](s) = (−1)n J (n) (s) where J(s) = L[j(t)](s) .

L[eat j(t)](s) = J(s − a) where J(s) = L[j(t)](s) .

L[u(t − c)j(t − c)](s) = e−cs J(s) where J(s) = L[j(t)](s), c ≥ 0,

and u is the unit step function.
L[δ(t − c)j(t)](s) = e−cs j(c) where c ≥ 0 and δ is the unit impulse.