Empirical runtime distributions on separabale functions (f1: Sphere, f2: Ellipsoid, f3: Rastrigin, f4: Büche-Rastrigin, f5: Linear Slope) with target values in {100, …, 1e-8} in dimension 20.

The following algorithms perform particularly well up to their individual maximum number of function evaluations: NEWUOA (NEW Unconstraint Optimization Algorithm builds a second order model using 2n+1 points and with minimal Frobenius norm), LSfminbnd (Axis-parallel line search with MATLAB fminbnd univariate search) and LSstep (Axis-parallel line search with the univariate STEP Select The Easiest Point, based on interval division).

higlighted_results_separable_functions.txt · Last modified: 2011/11/03 14:13 by vhm
CC Attribution-Noncommercial-Share Alike 3.0 Unported
Valid CSS Driven by DokuWiki Recent changes RSS feed Valid XHTML 1.0