Empirical runtime distributions on ill-conditioned functions (f10: High conditioning Ellipsoid, f11: Discus, f12: Bent Cigar, f13: Sharp Ridge, f14: Different Powers) 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: GLOBAL (Sampling, clustering and local search using BFGS or Nelder-Mead), iAMALGAM (Adapted Maximum-Likelihood Gaussian Model Iterated Density Estimation Algorithm with no-improvement stretch, anticipated mean shift and interlaced restarts with one large or several small populations with incremental model building) and BIPOP-CMA-ES (CMA-ES restarted with budgets for small and large population size).

higlighted_results_ill-conditioned_functions.txt · Last modified: 2011/11/03 14:17 by vhm
CC Attribution-Noncommercial-Share Alike 3.0 Unported
Valid CSS Driven by DokuWiki Recent changes RSS feed Valid XHTML 1.0