Abstract
Evolution Strategies (ESs) are population-based methods well suited for parallelization. In this paper, we study the convergence of the (μ/μ w ,λ)-ES, an ES with weighted recombination, and derive its optimal convergence rate and optimal μ especially for large population sizes. First, we theoretically prove the log-linear convergence of the algorithm using a scale-invariant adaptation rule for the step-size and minimizing spherical objective functions and identify its convergence rate as the expectation of an underlying random variable. Then, using Monte-Carlo computations of the convergence rate in the case of equal weights, we derive optimal values for μ that we compare with previously proposed rules. Our numerical computations show also a dependency of the optimal convergence rate in ln (λ) in agreement with previous theoretical results.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Preview
Unable to display preview. Download preview PDF.
References
Schumer, M., Steiglitz, K.: Adaptive step size random search. IEEE Transactions on Automatic Control 13, 270–276 (1968)
Rechenberg, I.: Evolutionstrategie: Optimierung Technisher Systeme nach Prinzipien des Biologischen Evolution. Fromman-Hozlboog Verlag, Stuttgart (1973)
Schwefel, H.-P.: Collective phenomena in evolutionary systems. In: Checkland, P., Kiss, I. (eds.) Problems of Constancy and Change-The Complementarity of Systems Approaches to Complexity, Proc. of 31st Annual Meeting Int’l Soc. for General System Research, Budapest, vol. 2, pp. 1025–1033 (1987)
Hansen, N., Ostermeier, A.: Completely derandomized self-adaptation in evolution strategies. Evolutionary Computation 9(2), 159–195 (2001)
Auger, A., Hansen, N.: Reconsidering the progress rate theory for evolution strategies in finite dimensions. In: ACM Press (ed.) Proceedings of the Genetic and Evolutionary Computation Conference (GECCO 2006), pp. 445–452 (2006)
Arnold, D.V.: Optimal weighted recombination. In: Foundations of Genetic Algorithms, vol. 8, pp. 215–237. Springer, Heidelberg (2005)
Jebalia, M., Auger, A., Liardet, P.: Log-linear convergence and optimal bounds for the (1+1)-ES. In: Monmarché, N., Talbi, E.-G., Collet, P., Schoenauer, M., Lutton, E. (eds.) EA 2007. LNCS, vol. 4926, pp. 207–218. Springer, Heidelberg (2008)
Teytaud, F., Teytaud, O.: On the parallel speed-up of Estimation of Multivariate Normal Algorithm and Evolution Strategies. In: Proceedings of EvoStar 2009, pp. 655–664 (2009)
Teytaud, F.: A new selection ratio for large population sizes. In: Proceedings of EvoStar 2010 (2010)
Bienvenüe, A., François, O.: Global convergence for evolution strategies in spherical problems: some simple proofs and difficulties. Th. Comp. Sc. 306(1-3), 269–289 (2003)
Jebalia, M., Auger, A., Hansen, N.: Log-linear convergence and divergence of the scale-invariant (1+1)-ES in noisy environments. Algorithmica (to appear 2010)
Jebalia, M., Auger, A.: Log-linear Convergence of the Scale-invariant (μ/μ w ,λ)-ES and Optimal μ for Intermediate Recombination for Large Population Sizes. Research Report n=°7275, INRIA (2010)
Auger, A.: Convergence results for (1,λ)-SA-ES using the theory of ϕ-irreducible markov chains. Theoretical Computer Science 334(1-3), 35–69 (2005)
Beyer, H.-G.: The Theory of Evolution Strategies. Nat. Comp. Series. Springer, Heidelberg (2001)
Beyer, H.-G., Sendhoff, B.: Covariance Matrix Adaptation revisited - The CMSA Evolution Strategy. In: Rudolph, G., Jansen, T., Lucas, S., Poloni, C., Beume, N. (eds.) PPSN 2008. LNCS, vol. 5199, pp. 123–132. Springer, Heidelberg (2008)
Teytaud, O., Fournier, H.: Lower Bounds for Evolution Strategies Using VC-dimension. In: Rudolph, G., et al. (eds.) Proceedings of PPSN X, pp. 102–111. Springer, Heidelberg (2008)
Beyer, H.-G.: Toward a Theory of Evolution Strategies: On the Benefits of Sex - the (μ/μ,λ) Theory. Evolutionary Computation 3(1), 81–111 (1995)
Author information
Authors and Affiliations
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2010 Springer-Verlag Berlin Heidelberg
About this paper
Cite this paper
Jebalia, M., Auger, A. (2010). Log-Linear Convergence of the Scale-Invariant (μ/μ w ,λ)-ES and Optimal μ for Intermediate Recombination for Large Population Sizes. In: Schaefer, R., Cotta, C., Kołodziej, J., Rudolph, G. (eds) Parallel Problem Solving from Nature, PPSN XI. PPSN 2010. Lecture Notes in Computer Science, vol 6238. Springer, Berlin, Heidelberg. https://doi.org/10.1007/978-3-642-15844-5_6
Download citation
DOI: https://doi.org/10.1007/978-3-642-15844-5_6
Publisher Name: Springer, Berlin, Heidelberg
Print ISBN: 978-3-642-15843-8
Online ISBN: 978-3-642-15844-5
eBook Packages: Computer ScienceComputer Science (R0)