 |
|
|
|
| HAL: inria-00276854, version 5 |
| arXiv: 0805.0231 |
| See detailed view |  BibTeX,EndNote,... |
|
|
| Available versions | v1 (2008-05-02) | v2 (2008-05-02) | v3 (2008-05-03) | v4 (2008-05-13) | v5 (2008-05-18) |
|
|
|
|
| CMA-ES with Two-Point Step-Size Adaptation |
|
|
Nikolaus Hansen 1, 2 |
|
|
| (2008) |
|
|
|
| We combine a refined version of two-point step-size adaptation with the covariance matrix adaptation evolution strategy (CMA-ES). Additionally, we suggest polished formulae for the learning rate of the covariance matrix and the recombination weights. In contrast to cumulative step-size adaptation or to the 1/5-th success rule, the refined two-point adaptation (TPA) does not rely on any internal model of optimality. In contrast to conventional self-adaptation, the TPA will achieve a better target step-size in particular with large populations. The disadvantage of TPA is that it relies on two additional objective function evaluations. |
|
|
|
|
|
|
|
|
|
|
|
| 1: | TAO (INRIA Futurs) |
|
|
|
INRIA РCNRS : UMR8623 РUniversit̩ Paris XI - Paris Sud |
|
|
| 2: | Microsoft Research - Inria Joint Centre (MSR - INRIA) |
|
|
|
INRIA – Microsoft – Microsoft Research Laboratory Cambridge |
|
|
|
|
|
|
|
|
|
| Domain | : | Computer Science/Neural and Evolutionary Computing |
|
|
| optimization – evolutionary algorithms – covariance matrix adaptation – step-size control – self-adaptation – two-point adaptation |
|
|
|
| Attached file list to this document: | ||||||||||
|
||||||||||
|
|
|
| inria-00276854, version 5 | |
| http://hal.inria.fr/inria-00276854 | |
| oai:hal.inria.fr:inria-00276854 | |
| From: Nikolaus Hansen | |
| Submitted on: Sunday, 18 May 2008 01:12:24 | |
| Updated on: Sunday, 18 May 2008 08:44:16 | |
