Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

I'm making an exercise in a book from Ravi P. Agarwal, Introduction to Ordinary Differential Equations. I would like to get some help for the following problem:

We are given the differential equation $$ y''+p_1(x)y'+p_2(x)y = 0 $$ where $p_1(x)$ and $p_2(x)$ are continuous functions, and periodic with period $w$ in $\mathbb{R}$.

Show that:

A nontrivial solution $y(x)$ is periodic of period $w$ if and only if $y(0)=y(w)$ and $y'(0)=y'(w)$.

Further if $y_1(x)$ and $y_2(x)$ are two solutions such that $y_1 (0)= 1, y_1'(0)=0, y_2(0)=1, y'_2(0)=0$ then show that:

there exist constants $a, b, c, d$ such that for all $x$ :

$$y_1(x+w)= ay_1(x)+by_2(x)$$ $$y_2(x+w)= cy_1(x)+dy_2(x)$$

I think I have to use a theorem which is in the book. I will give the theorem:

Theorem: Let the matrix $A(x)$ and the function $b(x)$ be continuous and periodic of period $w$ in $\mathbb{R}$. Then the differential system $ u'= A(x)u + b(x)$ has a periodic solution $u(x)$ of period $w$ if and only if $u(0)=u(w)$

share|cite|improve this question
It should be the case that $y_2(0)=0$ and $y_2'(0)=1$? – Ragnar Jan 14 '13 at 13:25
And you also ask for $y'(0)= y'(w)$ in the first part? – Ragnar Jan 14 '13 at 14:58
@беркай Yes you are right. I have fixed the question. Could you please give answer the question? :-) – Applied mathematician Jan 14 '13 at 22:10
up vote 2 down vote accepted

If $y(x)=y(x+w)$, then it is immediate that $y(0)=y(w)$ and $y'(0)=y'(w)$.

For the converse, suppose $y(x)$ be a solution of the given differential equation satisfying $y(0)=y(w)$ and $y'(0)=y'(w)$. Let $v(x)=y(x+w)$. Then,

$$ \tag 1 y''(x+w) + p_1(x+w)y'(x+w) + p_2(x+w)y(x+w)=0 $$

is equivalent to

$$ v''(x) + p_1(x)v'(x) + p_2(x)v(x)=0. $$

Hence, $v(x)$ satisfies the given differential equation. Since, $v(0)=y(w)=y(0)$ and $v'(0)=y'(w)=y'(0)$, by the uniqueness of the initial value problems, it follows that $y(x)=v(x)=y(x+w)$, i.e. $y(x)$ is periodic of period $w$.

Since the Wronskian $W(y_1,y_2)(0)=1 \neq 0$, the solutions $y_1(x)$ and $y_2(x)$ are independent. Moreover, $y_1(x+w)$ and $y_2(x+w)$ satisfy $(1)$. Hence, both can be written as linear combinations of $y_1(x)$ and $y_2(x)$, i.e., $$ y_1(x+w)= a y_1(x) + by_2(x) ; \quad y_2(x+w)= c y_1(x) + d y_2(x). $$

Using the initial conditions, it follows that $$a=y_1(w); \quad b=y_1'(w);\quad c=y_2(w);\quad d=y_2'(w).$$

share|cite|improve this answer
Great! Thank you for your help. Can you look at the second part of my question? – Applied mathematician Jan 15 '13 at 0:22
By the way, you should edit the statement: $y_2(0)=0$ and $y_2'(0)=1$. – Ragnar Jan 15 '13 at 0:43
The Wronskian at the point $0$ must be $1$, this is the idea. We have $W(x)=y_1(x)y_2'(x)-y_1'(x)y_2(x)$ and $W(0)=1$ doesn't hold with your assumption. – Ragnar Jan 15 '13 at 0:57
I mean we have the Wronskian $1$ with the fundamental solutions satisfying $y_1(0)=1$, $y_1'(0)=0$ etc. Also, the Existence-Uniqueness Theorem assures the existence of these solutions under some conditions, but we are already given these solutions so what I wrote was misleading and I removed this part. – Ragnar Jan 15 '13 at 1:14
Ok! I will edit my answer. – Ragnar Jan 15 '13 at 1:23

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.