$f,g:[0,1 ]\to [0,1]$ be continuous functions and twice differentiable in $[0,1]$ such that $g'(x) \ne 0 ,\forall x \in (0,1) , f''(0)g'(0) \ne f'(0)g''(0) $ , let $ \theta(x)$ be one of the numbers for which the assertion of the Cauchy's generalized mean-value theorem holds i.e. $\dfrac {f(x)-f(0)}{g(x)-g(0)}=\dfrac {f'\big( \theta(x)\big)}{g'\big( \theta(x)\big)}$ , then how to compute $\lim_{x \to 0+} \dfrac{\theta(x)}x$ ?
2 Answers
Let $x>0$. There exists $c_x,d_x,a_x,b_x \in (0,x)$ such that:
$$f(x)=f(0)+xf^{\prime}(0)+\frac{x^2}{2}f^{\prime\prime}(c_x)$$
$$g(x)=g(0)+xg^{\prime}(0)+\frac{x^2}{2}g^{\prime\prime}(d_x)$$
$$f^{\prime}(\theta(x))=f^{\prime}(0)+\theta(x)f^{\prime\prime}(a_x)$$
and
$$g^{\prime}(\theta(x))=f^{\prime}(0)+\theta(x)f^{\prime\prime}(b_x)$$
Now from your equality, you get:
$$\theta(x)\big [(f^{\prime}(0)g^{\prime\prime}(b_x)-g^{\prime}(0)f^{\prime\prime}(a_x))+\frac{x}{2}(f^{\prime\prime}(c_x)g^{\prime\prime}(b_x)-f^{\prime\prime}(a_x)g^{\prime\prime}(d_x))\big ]=\frac{x}{2}(f^{\prime}(0)g^{\prime\prime}(d_x)-g^{\prime}(0)f^{\prime\prime}(c_x))$$
Now as $x \to 0$:
$$\frac{x}{2}(f^{\prime\prime}(c_x)g^{\prime\prime}(b_x)-f^{\prime\prime}(a_x)g^{\prime\prime}(d_x)) \to 0$$
$$f^{\prime}(0)g^{\prime\prime}(b_x)-g^{\prime}(0)f^{\prime\prime}(a_x)\to f^{\prime}(0)g^{\prime\prime}(0)-g^{\prime}(0)f^{\prime\prime}(0)=L \not =0$$
and $$f^{\prime}(0)g^{\prime\prime}(d_x)-g^{\prime}(0)f^{\prime\prime}(c_x))\to L$$ and so $\displaystyle \frac{\theta(x)}{x}\to \frac{1}{2}$.
- 18,734
- 26
- 36
-
how did you get $\theta(x)\big [(f^{\prime}(0)g^{\prime\prime}(b_x)-g^{\prime}(0)f^{\prime\prime}(a_x))+\frac{x}{2}(f^{\prime\prime}(c_x)g^{\prime\prime}(b_x)-f^{\prime\prime}(a_x)g^{\prime\prime}(d_x))\big ]=\frac{x}{2}(f^{\prime}(0)g^{\prime\prime}(d_x)-g^{\prime}(0)f^{\prime\prime}(c_x))$ ? – Souvik Dey Jul 09 '14 at 13:45
-
You have $$\dfrac {f(x)-f(0)}{g(x)-g(0)}=\dfrac {f'\big( \theta(x)\big)}{g'\big( \theta(x)\big)}$$. Use the formulae I have given, Simplify the first term by $x$, and then multiply by the denominators. – Kelenner Jul 09 '14 at 13:51
More observation than solution: If you let $f(x)=ax^2+bx$ and $g(x)=cx^2+dx$, then, in general, the conditions of the problem are satisfied, and the equation becomes
$${ax^2+bx\over cx^x+dx}={ax+b\over cx+d}={2a\theta+b\over2c\theta+d}$$
which solves to
$${\theta\over x}={1\over2}$$
This accords with the answer just given by Kelenner.
(Note: the OP's comment and my reply were to an earlier version of the answer, which failed to notice the condition $f''(0)g'(0)\not=f'(0)g''(0)$.)
- 79,832
-
but the functions you have taken doesn't satisfy $f''(0)g'(0) \ne f'(0)g''(0)$ – Souvik Dey Jul 09 '14 at 13:03
-
1