Abstract
It is a known issue that Estimation of Distribution Algorithms (EDAs) tend to converge prematurely. There exist various articles which report that they are incapable of determining the adequate variance, getting trapped even in regions which does not contain any global neither local minimum. In this vein, several proposals intend to insert and to preserve diversity. The proposal range is from modifying the probability distribution by inserting artificial variance, to mutating individuals or the search distribution. This work presents a novel selection method which is used to estimate a weighted covariance matrix equipped with adequate magnitude and directions. The covariance matrix is directly related with the information acquired from the current population, hence no artificial artifacts are inserted. By comparing with state of the art EDAs using test problems and graphical and statistical measures, we evidence that our proposal avoids premature convergence, and solves difficult problems -in the sense of variance preservation-, without the need of complex models, mutation, or explicit variance scaling.
This is a preview of subscription content, log in via an institution.
Buying options
Tax calculation will be finalised at checkout
Purchases are for personal use only
Learn about institutional subscriptionsPreview
Unable to display preview. Download preview PDF.
References
Bosman, P.A., Thierens, D.: Advancing continuous ideas with mixture distributions and factorization selection metrics. In: Proceedings of the Optimization by Building and Using Probabilistic Models OBUPM Workshop at the Genetic and Evolutionary Computation Conference GECCO 2001, pp. 208–212. Morgan Kaufmann (2001)
Bosman, P.A.N., Grahl, J., Rothlauf, F.: SDR: A Better Trigger for Adaptive Variance Scaling in Normal EDAs. In: GECCO 2007: Proceedings of the 8th Annual Conference on Genetic and Evolutionary Computation, pp. 516–522. ACM (2007)
Cai, Y., Sun, X., Xu, H., Jia, P.: Cross entropy and adaptive variance scaling in continuous eda. In: Proceedings of the 9th Annual Conference on Genetic and Evolutionary Computation, GECCO 2007, pp. 609–616. ACM, New York (2007), doi:10.1145/1276958.1277081
Cho, D.-Y., Zhang, B.-T.: Evolutionary Continuous Optimization by Distribution Estimation with Variational Bayesian Independent Component Analyzers Mixture Model. In: Yao, X., Burke, E.K., Lozano, J.A., Smith, J., Merelo-Guervós, J.J., Bullinaria, J.A., Rowe, J.E., Tiňo, P., Kabán, A., Schwefel, H.-P. (eds.) PPSN VIII 2004. LNCS, vol. 3242, pp. 212–221. Springer, Heidelberg (2004)
Dong, W., Yao, X.: NichingEDA: Utilizing the diversity inside a population of EDAs for continuous optimization. In: IEEE Congress on Evolutionary Computation, pp. 1260–1267 (2008)
Dong, W., Yao, X.: Unified eigen analysis on multivariate gaussian based estimation of distribution algorithms. Information Sciences 178(15), 3000–3023 (2008), doi:10.1016/j.ins.2008.01.021
Grahl, J., Bosman, P.A.N., Minner, S.: Convergence Phases, Variance Trajectories, and Runtime Analysis of Continuos EDAs. In: GECCO 2007: Proceedings of the 9th Annual Conference on Genetic and Evolutionary Computation, pp. 516–522. ACM (2007)
Grahl, J., Bosman, P.A.N., Rothlauf, F.: The correlation-triggered adaptive variance scaling idea. In: GECCO 2006: Proceedings of the 8th Annual Conference on Genetic and Evolutionary Computation, pp. 397–404. ACM Press (2006)
Handa, H.: The effectiveness of mutation operation in the case of estimation of distribution algorithms. Biosystems 87(23), 243–251 (2007)
Larrañaga, P., Lozano, J.A.: Estimation of Distribution Algorithms: A New Tool for Evolutionary Computation. Kluwer Academic Publishers, Norwell (2001)
Ocenasek, J., Schwarz, J.: Estimation of distribution algorithm for mixed continuous-discrete optimization problems. In: 2nd Euro-International Symposium on Computational Intelligence, pp. 227–232. IOS Press (2002)
Shapiro, J.L.: Diversity Loss in General Estimation of Distribution Algorithms. In: Runarsson, T.P., Beyer, H.-G., Burke, E.K., Merelo-Guervós, J.J., Whitley, L.D., Yao, X. (eds.) PPSN 2006. LNCS, vol. 4193, pp. 92–101. Springer, Heidelberg (2006)
Teytaud, F., Teytaud, O.: Why one must use reweighting in estimation of distribution algorithms. In: Proceedings of the 11th Annual Conference on Genetic and Evolutionary Computation, GECCO 2009, pp. 453–460. ACM, New York (2009), http://doi.acm.org/10.1145/1569901.1569964 , doi: http://doi.acm.org/10.1145/1569901.1569964
Valdez, S.I., Hernández, A., Botello, S.: Efficient Estimation of Distribution Algorithms by Using the Empirical Selection Distribution. In: Korosec, P. (ed.) New Achievements in Evolutionary Computation. InTech (2010)
Yuan, B., Gallagher, M.: On the importance of diversity maintenance in estimation of distribution algorithms. In: Proceedings of the 2005 Conference on Genetic and Evolutionary Computation, GECCO 2005, pp. 719–726. ACM, New York (2005), http://doi.acm.org/10.1145/1068009.1068129 , doi: http://doi.acm.org/10.1145/1068009.1068129
Zhang, Q., Mühlenbein, H.: On the Convergence of a Class of Estimation of Distribution Algorithms. IEEE Transactions on Evolutionary Computation 8(2), 127–136 (2004)
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2013 Springer-Verlag Berlin Heidelberg
About this paper
Cite this paper
Ivvan Valdez, P.S., Hernández-Aguirre, A., Botello, S. (2013). Adequate Variance Maintenance in a Normal EDA via the Potential-Selection Method. In: Schütze, O., et al. EVOLVE - A Bridge between Probability, Set Oriented Numerics, and Evolutionary Computation II. Advances in Intelligent Systems and Computing, vol 175. Springer, Berlin, Heidelberg. https://doi.org/10.1007/978-3-642-31519-0_14
Download citation
DOI: https://doi.org/10.1007/978-3-642-31519-0_14
Publisher Name: Springer, Berlin, Heidelberg
Print ISBN: 978-3-642-31518-3
Online ISBN: 978-3-642-31519-0
eBook Packages: EngineeringEngineering (R0)