by Tianjun Liao, Marco A. Montes de Oca, and Thomas Stützle
April 2011
| Submitted to Soft Computing Journal |
In this article, we apply an automatic algorithm
configuration tool to improve the performance of the CMA-ES algorithm with increasing population size (iCMA-ES),
the best performing algorithm on the CEC'05 benchmark
set for continuous function optimization. In particular, we
consider a separation between tuning and test sets and, thus,
tune iCMA-ES on a different set of functions than the ones
of the CEC'05 benchmark set. Our experimental results show
that the tuned iCMA-ES improves significantly over the default version of iCMA-ES. Furthermore, we provide some
further analyses on the impact of the modified parameter settings on iCMA-ES performance and a comparison to recent
results of algorithms that use CMA-ES as a subordinate local search.
Please look into the attached supplementary.pdf
1. The qualied run-length distributions (RLDs, for short) over 100 independent runs obtained by iCMA-ES-dp and iCMA-ES-tsc on the 50 dimensional versions of functions fcec4, fcec5, fcec11, fcec12, fcec16 and fcec17.
2. The average errors obtained by Sep-iCMA-ES-tsc and iCMA-ES-tsc over 25 independent runs for CEC'05 functions.
3. The data obtained by different variations of the tuning setups over 25 independent runs for CEC'05 functions of dimension 50.