Define $f \in C^{2}[a,b]$ satisfying $f''+pf'+qf=0$ with $f(a)=f(b)=0$, where $p\in C^{0}[a,b]$ and $q\in C^{0}[a,b]$ are two functions.

If $q\leq0$, can we prove $f\equiv 0$ $\forall x\in [a,b]$ ?

My try:
If $f\not\equiv 0$, without loss of generality, we assume that the maximum of $f$ on $[a,b]$ is greater than zero, while notating $f(x_0)=\displaystyle\max_{[a,b]} f$.

Then we have $f(x_0) > 0$, $f'(x_0) = 0$ and $f''(x_0) \leq 0$.

I figured out that if we alter the condition $q\leq0$ into $q(x)<0$ there evidently exists contradiction.

But how to analyze further with the condition $q\leq0$? Can we still find contradiction if $q(x_0)=0$ and $f''(x_0)=0$ ?

Any ideas would be highy appreciated!


The following is based on the classical proof of the maximum principle.

Let $L(f)=f''+p\,f'+q\,f$. If $L(f)>0$, then your argument shows that $f$ must be identically $0$.

But we have $L(f)=0$, not $>0$. What can we do? Take $M>0$ such that $M^2+M\,p(x)+q(x)>0$ for all $x\in[a,b]$ and let $\epsilon>0$. Then $$ L(f+\epsilon\,e^{Mx})=\epsilon\,e^{Mx}(M^2+M\,p(x)+q(x))>0\quad\forall x\in[a,b]. $$ Then $$ \max_{a\le x\le b}(f+\epsilon\,e^{Mx})=\max\bigl(f(a)+\epsilon\,e^{Ma},f(b)+\epsilon\,e^{Mb}\bigr)=\epsilon\,e^{Mb}. $$ Letting $\epsilon\to0$ gives the desired result.

  • $\begingroup$ Your answer does help! I got lost in finding contradiction by analyzing $f(x_0 +s)$ ($s\to 0$), which is not as brilliant as your answer is. $\endgroup$ – Zero Nov 28 '18 at 16:06
  • $\begingroup$ All I did was an adaptation of the proof of the maximum principle. $\endgroup$ – Julián Aguirre Nov 28 '18 at 17:33

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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