4

Assume on the contrary, $f'(x)\geq 0$, and since $f''(x)\leq 0$, then $f(x)$ is concave down and decreasing, so $f(x)$ will eventually less than $0$, contradict to $f(x)>0$, but how can I prove that $f(x)$ will eventually less than $0$ formally? thanks!

  • Suppose $f'(a) < 0$ for some $a > 0$. Define $g(x)$ in the following way. $g(x) = f(x)$ when $x \leq a$ and $g''(x) = 0$ when $x > a$. Then you can show that,
    1. For sufficiently large $x$, $g(x) < 0$ (Hint: try explicitly writing $g(x)$ in terms of $f(a)$ when $x > a$).

    2. $f(x) \leq g(x)$ so that $f(x) < 0$ when $x$ is sufficiently large.

    – forgottenarrow Apr 25 '20 at 22:24
  • Small mistake in the hint above: write $g(x)$ in terms of $f(a)$ and $f'(a)$, not just $f(a)$. – forgottenarrow Apr 25 '20 at 22:30

1 Answers1

4

Suppose $f'(a) <0$. Since $f'$ is non-increasing we get $f'(x) \leq f'(a)$ for all $x \geq a$. Hence $0 <f(x)=f(a)+\int_a^{x} f'(t)dt \leq f(a)+f'(a)(x-a)$ for $x >a$. As $x \to \infty$ RHS tends to $-\infty$ leading to a contradiction.