论文信息 - Optimal mixing evolutionary algorithms

Optimal mixing evolutionary algorithms

A key search mechanism in Evolutionary Algorithms is the mixing or juxtaposing of partial solutions present in the parent solutions. In this paper we look at the efficiency of mixing in genetic algorithms (GAs) and estimation-of-distribution algorithms (EDAs). We compute the mixing probabilities of two partial solutions and discuss the effect of the covariance build-up in GAs and EDas. Moreover, we propose two new Evolutionary Algorithms that maximize the juxtaposing of the partial solutions present in the parents: the Recombinative Optimal Mixing Evolutionary Algorithm (ROMEA) and the Gene-pool Optimal Mixing Evolutionary Algorithm (GOMEA).

Dirk Thierens | Peter A. N. Bosman | P. Bosman | D. Thierens

[1] Dirk Thierens,et al. The Linkage Tree Genetic Algorithm , 2010, PPSN.

[2] Dirk Thierens,et al. Convergence Models of Genetic Algorithm Selection Schemes , 1994, PPSN.

[3] Kumara Sastry,et al. Linkage Learning via Probabilistic Modeling in the Extended Compact Genetic Algorithm (ECGA) , 2006, Scalable Optimization via Probabilistic Modeling.

[4] Dirk Thierens,et al. Pairwise and problem-specific distance metrics in the linkage tree genetic algorithm , 2011, GECCO '11.

[5] Martin Pelikan,et al. Scalable Optimization via Probabilistic Modeling: From Algorithms to Applications (Studies in Computational Intelligence) , 2006 .

[6] Martin V. Butz,et al. Performance of evolutionary algorithms on NK landscapes with nearest neighbor interactions and tunable overlap , 2009, GECCO '09.