IMC / 2015 / Problems / Day 2, P7
IMC 2015 · Day 2 · P7
hardCompute (Proposed by Jan Šustek, University of Ostrava)
Solution 1 of 2 (official)
We prove that For the integrand is greater than 1, so In order to find a tight upper bound, fix two real numbers, and , and split the interval into three parts at the points and . Notice that for sufficiently large (i.e., for with some ) we have .) For the integrand is decreasing, so we can estimate it by its value at the starting points of the intervals:
&= \frac{1}{A} \Bigl( \delta \cdot A + (K \log A - 1 - \delta) A^{\frac{1}{1+\delta}} + (A - K \log A) A^{\frac{1}{K \log A}} \Bigr) < \\ &< \frac{1}{A} \Bigl( \delta A + K A^{\frac{1}{1+\delta}} \log A + A \cdot A^{\frac{1}{K \log A}} \Bigr) = \delta + K A^{-\frac{\delta}{1+\delta}} \log A + e^{\frac{1}{K}}. \end{align*} Hence, for we have Taking the limit we obtain Now from and we get so and therefore
Solution 2 of 2 (official)
We will employ l'Hospital's rule.
Let , , and . Since and are continuous, the parametric integrals and are differentiable with respect to , and and Since , we can see that . Aplying l'Hospital's rule to we get so Now applying l'Hospital's rule to we get
How the field did
Score distribution (field cohort)
Computed on contestants with a meaningful total (field cohort); discrimination is the corrected item–total correlation.