Problem 1197. Numerical Integration
Please, add to the problem description that for values smaller than one, we should use 1 + x^2/2 for approximating the integrand. For instance, the answer provided by vpaintegral (Symbolic Math Toolbox; commented in my solution) is not accepted by the 4th test case: I believe that vpaintegral returns the exact solution*. The precision for the 4th test should be way smaller, 1e-2/1e-3, or the problem description needs to be changed.
* vpaintegral can handle values that cause the MATLAB integral function to overflow or underflow.
As a sanity check, I even did Monte Carlo Integration for the 4th case, and it still did not work, because the problem is not asking for the exact solution, but an approximation of an approximation (and that's why the requested precision should be smaller).
What is the point of this question? Everyone is just using built-in integral functions anyway. You cannot actually code a solution to this problem because the precision is way too high.
Is it cheating if the test suite appeals to the same approximation, claiming it to be "good enough"?
Problem Recent Solvers36
Find the Oldest Person in a Room
More from this Author7
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!Start Hunting!