inria-00287863, version 1
When does quasi-random work ?
Parallel Problem Solving from Nature (2008)
Abstract: We experiment the efficiency of quasi-random mutations in evolution strategies in continuous domains from various points of views: (i) non-convexity (ii) convergence rate (iii) non-asymptotic behavior (iv) noise. We conclude that quasi-random mutations are great.
- 1: Laboratoire de Recherche en Informatique (LRI)
- CNRS : UMR8623 – Université Paris XI - Paris Sud
- 2: TAO (INRIA Saclay - Ile de France)
- INRIA – CNRS : UMR8623 – Université Paris XI - Paris Sud
- Domain : Mathematics/Optimization and Control
- Keywords : Evolution strategies – derandomization – non-asymptotic results
- inria-00287863, version 1
- http://hal.inria.fr/inria-00287863
- oai:hal.inria.fr:inria-00287863
- From: Olivier Teytaud
- Submitted on: Friday, 13 June 2008 10:42:09
- Updated on: Thursday, 14 August 2008 11:48:35
Associated documents
Export