论文信息 - ARestart CMA Evolution Strategy WithIncreasing Population Size

ARestart CMA Evolution Strategy WithIncreasing Population Size

Inthis paperweintroduce arestart-CMA- evolution strategy, where thepopulation size isincreased foreachrestart (IPOP). Byincreasing thepopulation size thesearch characteristic becomes moreglobal af- tereachrestart. TheIPOP-CMA-ES isevaluated onthe test suit of25functions designed forthespecial session onreal-parameter optimization ofCEC2005. Itsperfor- manceiscompared toalocal restart strategy withcon- stant small population size. Onunimodal functions the performance issimilar. Onmulti-modal functions the local restart strategy significantly outperforms IPOPin 4test cases whereas IPOPperforms significantly better in29outof60tested cases.

Nikolaus Hansen | N. Hansen

[1] Nikolaus Hansen,et al. Adapting arbitrary normal mutation distributions in evolution strategies: the covariance matrix adaptation , 1996, Proceedings of IEEE International Conference on Evolutionary Computation.

[2] Jing J. Liang,et al. Problem Definitions and Evaluation Criteria for the CEC 2005 Special Session on Real-Parameter Optimization , 2005 .

[3] Nikolaus Hansen,et al. Completely Derandomized Self-Adaptation in Evolution Strategies , 2001, Evolutionary Computation.

[4] Nikolaus Hansen,et al. Invariance, Self-Adaptation and Correlated Mutations and Evolution Strategies , 2000, PPSN.

[5] Petros Koumoutsakos,et al. Reducing the Time Complexity of the Derandomized Evolution Strategy with Covariance Matrix Adaptation (CMA-ES) , 2003, Evolutionary Computation.