A Generalization of the CMA-ES Algorithm for Functions with Matrix Input

Simon Konzett

2014

Abstract

This paper proposes a novel modification to the covariance matrix adaptation evolution strategy (CMA-ES) introduced by (Hansen and Ostermeier, 1996) under a special problem setting. In this paper the case is considered when the function which has to be optimized takes a matrix as input. Here an approach is presented where without vectorizing directly matrices are sampled and column and row-wise covariance matrices are adapted in each iteration of the proposed evolution strategy to adapt the mutation distribution. The method seems to be able to capture correlations in the entries of the considered matrix and adapt the corresponding covariance matrices accordingly. Numerical tests are performed on the proposed method to show advantages and disadvantages.

References

  1. Dawid, A. P. (1981). Some matrix-variate distribution theory: notational considerations and a Bayesian application. Biometrika, 68(1):265-274.
  2. Hansen, N. (1998). Verallgemeinerte individuelle Schrittweitenregelung in der Evolutionsstrategie. Mensch & Buch Verlag, Berlin.
  3. Hansen, N. and Ostermeier, A. (1996). Adapting arbitrary normal mutation distributions in evolution strategies: The covariance matrix adaptation. In Evolutionary Computation, 1996., Proceedings of IEEE International Conference on, pages 312-317. IEEE.
  4. Hansen, N. and Ostermeier, A. (2001). Completely derandomized self-adaptation in evolution strategies. Evolutionary Computation, 9(2):159-195.
  5. Igel, C., Suttorp, T., and Hansen, N. (2006). A computational efficient covariance matrix update and a (1+ 1)-CMA for evolution strategies. In Proceedings of the 8th annual conference on Genetic and evolutionary computation, pages 453-460. ACM.
  6. Igeland, C., Hansen, N., and Roth, S. (2007). Covariance matrix adaptation for multi-objective optimization. Evolutionary computation, 15(1):1-28.
  7. Jastrebski, G. and Arnold, D. (2006). Improving evolution strategies through active covariance matrix adaptation. In Evolutionary Computation, 2006. CEC 2006. IEEE Congress on, pages 2814-2821. IEEE.
  8. N. Hansen (2011). The CMA Evolution Strategy: A Tutorial.
  9. Schwefel, H. (1977). Numerische Optimierung von Computer-Modellen mittels der Evolutionsstrategie. Birkhäuser Verlag, Basel.
Download


Paper Citation


in Harvard Style

Konzett S. (2014). A Generalization of the CMA-ES Algorithm for Functions with Matrix Input . In Proceedings of the International Conference on Evolutionary Computation Theory and Applications - Volume 1: ECTA, (IJCCI 2014) ISBN 978-989-758-052-9, pages 337-342. DOI: 10.5220/0005159703370342


in Bibtex Style

@conference{ecta14,
author={Simon Konzett},
title={A Generalization of the CMA-ES Algorithm for Functions with Matrix Input},
booktitle={Proceedings of the International Conference on Evolutionary Computation Theory and Applications - Volume 1: ECTA, (IJCCI 2014)},
year={2014},
pages={337-342},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0005159703370342},
isbn={978-989-758-052-9},
}


in EndNote Style

TY - CONF
JO - Proceedings of the International Conference on Evolutionary Computation Theory and Applications - Volume 1: ECTA, (IJCCI 2014)
TI - A Generalization of the CMA-ES Algorithm for Functions with Matrix Input
SN - 978-989-758-052-9
AU - Konzett S.
PY - 2014
SP - 337
EP - 342
DO - 10.5220/0005159703370342