Take the 2-minute tour ×
Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. It's 100% free, no registration required.

Suppose $$ F(a)=0, F(b)=0, a<b $$ and $F''(x)<0$ for $a<x<b$.

Do we have $F(x) > 0$ for $a<x<b$?

Can you tell me an existing theorem that I can use in a paper to obtain the conclusion? I think there should be one, but I can't remember.

share|improve this question
    
If $F(x) < 0$ for some $a < x < b$, then you can use the mean value theorem three times to show that $F''(X) > 0$ somewhere. –  arkeet Dec 11 '12 at 7:51
    
@arkeet I mean, can you tell me an existing theorem that I can use in a paper to obtain the conclusion? I think there should be one, but I can't remember. –  Chen Dec 11 '12 at 8:14
add comment

3 Answers 3

This is a consequence of the convexity of $-F(x)$ on $[a,b]$.

Suppose $F(c) < 0$ for some $c \in (a,b)$. By the mean value theorem, there exist values $d_1 \in (a,c)$ and $d_2 \in (c,b)$ such that \begin{align} F'(d_1) = \frac{F(c) - F(a)}{c - a} = \frac{F(c)}{b-a} < 0 \\ ~ \\ F'(d_2) = \frac{F(b) - F(c)}{b - c} = \frac{-F(c)}{c-b}> 0, \end{align} and hence $F'(x)$ changes sign somewhere on $(a,b)$. This contradicts $F''(x) < 0$. Hence, $F(c) \geq 0$ for all $c \in (a,b)$.

share|improve this answer
add comment

There is no theorem that I know of that talks specifically about this case, but you can prove it using a bit of simple calculus.

How about this proof: By Rolle's theorem, $F'(x) = 0$ at least once in the interval $(a, b)$. However it must happen exactly once since $F''(x) < 0$... This critical point is a maximum point, so we have $F(x)$ ascending from $(a, 0)$ towards a maximum and then descending back to $(b, 0)$.

share|improve this answer
add comment

Why not. can be concave/not and positive between any two points. the concavity of the curve may also be changing at x. Depends on the function.

share|improve this answer
add comment

Your Answer

 
discard

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.