Math6340 Ordinary Differential Equations
Assignment I
Due Date: February 28, 2021
1.
Use at least two methods
(including Putzer Algorithm) to find etA for A =
0 −2 −1 −1
1 2 1 1
.
0 1 1 0
0 0 0 0
2. Let A be a constant n × n matrix.
(1). Show that deteA = eT rA .
(2). How should α be chosen so that lim e−αt eAt = 0?
t→∞
3. Consider a mass-spring system y 00 + ky 0 + y = 0, k is a non-negative parameter.
(a) Transform to a system by introducing v = y 0 as velocity.
(b) Plot the T D plane and different phase portraits using k as parameter.
For each case of k, plot in one figure for y, v over time t.
(c) For k = 1, find the general solution matrix of the system.
(d) Consider y 00 + y 0 + y = f (t), f (t) is bounded for t ≥ 0, discuss the limit
lim y(t).
t→∞
4. Suppose φ1 (t) and φ2 (t) are two continuously differentiable functions satisfying
|φ1 (0) − φ2 (0)| < δ
|φ01 (t) − φ02 (t)| ≤ M |φ1 (t) − φ2 (t)| + h(t), (0 ≤ t ≤ T )
for some constants δ, k > 0 and continuous nonnegative function h(t). Show
that Z t
Mt
|φ1 (t) − φ2 (t)| ≤ δe + eM (t−s) h(s)ds, (0 ≤ t ≤ T ).
0
5. Consider the initial value problem x0 = x2 , x(0) = 1.
(a) Solve the initial value problem.
(b) If you apply the Picard’s existence and uniqueness Theorem, what is the
maximum interval of existence you can get? Justify your answer.
(c) Construct the Picard’s successive approximation xi (t) for i = 0, 1, 2, ..., n
and verify that xn (t) = Σnj=0 tj . If x(t) is the unique solution, then show
that |x(t) − xn (t)| = O(tn ), t ∈ (−1, 1).
6. Assume that f (x) ∈ C(R). Show that for any t0 , x0 ∈ R, the initial value
problem dx
dt
= f 2 (x) + et , x(t0 ) = x0 has a unique solution.
1
7. Assume that f (x) ∈ C(R) and f (0) = 0, f (x) > 0, x 6= 0. Prove that
x0 = f (x), x(0) = 0
RC
has a unique solution if and only if 0 fdx(x)
= ∞ for any C ∈ R.
8. Consider x0 = f (t, x) where f (t, x) is defined for (t, x) ∈ [0, 1] × R by
f (0, x) =0
f (t, x) = 2t if x < 0,
4x
f (t, x) = 2t − t
if 0 ≤ x ≤ t2 , t > 0,
f (t, x) = −2t if x > t2 .
Verify that f is continuous, but it does not satisfy a Lipschitz condition.
For the initial point (t0 , x0 ) = (0, 0), show that the successive approximation
does not converge for t ∈ I, and none of its subsequences converges to a
solution, although the solution is indeed unique.
d
Hint: For uniqueness, show that |φ
dt 1
− φ2 | ≤ 0 for any two solutions φ1 , φ2 , by using f is monotone in
x for each fixed t. This example appears in the book of Coddington-Levinson page 53.
9. Prove the generalized Gronwall’s inequality.
If φ, α are real valued and continuous for t ∈ [a, b], β(t) > 0 is integrable on
[a, b] and Z t
φ(t) ≤ α(t) + β(s)φ(s)ds, a ≤ t ≤ b,
a
then
Z t Z t
φ(t) ≤ α(t) + α(s)β(s) exp β(u)du ds, a ≤ t ≤ b.
a s
Explain that Gronwall’s1 inequality is a special case if α is continuous with its
first derivative nonnegative.
10. Let
x sin x1
x 6= 0
f1 (x) =
0 x=0
and
x2 sin x1
x 6= 0
f2 (x) =
0 x=0
Let I = (a, b) with a < 0 < b.
(a) For each function fi (x) (i = 1, 2), does it satisfy a Lipschitz condition on
I? Justify your answer.
(b) For i = 1 and 2, is the solution to the initial value problem dx dt
=
fi (x), x(0) = 0 unique? Justify your answer.
1
Gronwall’s Inequality first appeared in Ann. of Math. 20 (1919), 292-296, 4, T. H. Gronwall