I packed CMA-ES in 2014 for Scilab 5.3, but it didn’t work for Scilab 6. Thank you to Stephane Mottelet and Clement David for helping to update CMA-ES so it now runs under Scilab 6, and for helping to package a module properly.
CMA-ES is the most powerful black-box optimization routine available, as proved by many years of dominance in the BBOB competitions. Originally so-called genetic algorithms provided smart black-box optimization, and later optimization routines typically had a decade of being ‘best’ like particle swarm optimization, and simulated annealing, etc. CMA-ES seems to the so powerful, nobody was able to formulate a better general-purpose black-box optimization routine.
I tried to access the link (in ATOMS module’s description) about intro to CMA -ES but it seems is broken…
I found in HAL for LRI about CMA ES from Nikolaus Hansen. The CMA Evolution Strategy: A Tutorial. 2005. ⟨hal-01297037v2⟩
and Oswin Krause, Tobias Glasmachers, Nikolaus Hansen, Christian Igel. Unbounded Population MO-CMA-ES for the Bi-Objective BBOB Test Suite. GECCO’16 - Companion of Proceedings of the 2016 Genetic and Evolutionary Computation Conference, ACM, Jul 2016, Denver, United States. pp.1177-1184, ⟨10.1145/2908961.2931699⟩. ⟨hal-01381653⟩