Oversized Populations and Cooperative Selection: Dealing with Massive Resources in Parallel Infrastructures

This paper proposes a new selection scheme for Evolutionary Algorithms EAs based on altruistic cooperation between individuals. Cooperation takes place every time an individual undergoes selection: the individual decreases its own fitness in order to improve the mating chances of worse individuals. On the one hand, the selection scheme guarantees that the genetic material of fitter individuals passes to subsequent generations as to decrease their fitnesses individuals have to be firstly selected. On the other hand, the scheme restricts the number of times an individual can be selected not to take over the entire population. We conduct an empirical study for a parallel EA version where cooperative selection scheme is shown to outperform binary tournament: both selection schemes yield the same qualities of solutions but cooperative selection always improves the times to solutions.

[1]  Dirk Sudholt,et al.  General Scheme for Analyzing Running Times of Parallel Evolutionary Algorithms , 2010, PPSN.

[2]  Kenneth A. De Jong,et al.  A Cooperative Coevolutionary Approach to Function Optimization , 1994, PPSN.

[3]  Zbigniew Michalewicz,et al.  Parameter Setting in Evolutionary Algorithms , 2007, Studies in Computational Intelligence.

[4]  Enrique Alba,et al.  The exploration/exploitation tradeoff in dynamic cellular genetic algorithms , 2005, IEEE Transactions on Evolutionary Computation.

[5]  Martijn C. Schut,et al.  Boosting Genetic Algorithms with Self-Adaptive Selection , 2006, 2006 IEEE International Conference on Evolutionary Computation.

[6]  Juan Julián Merelo Guervós,et al.  EvAg: a scalable peer-to-peer evolutionary algorithm , 2010, Genetic Programming and Evolvable Machines.

[7]  A. E. Eiben,et al.  Introduction to Evolutionary Computing , 2003, Natural Computing Series.

[8]  Juan Julián Merelo Guervós,et al.  Characterizing Fault-Tolerance of Genetic Algorithms in Desktop Grid Systems , 2010, EvoCOP.

[9]  Cláudio F. Lima,et al.  Adaptive Population Sizing Schemes in Genetic Algorithms , 2007, Parameter Setting in Evolutionary Algorithms.

[10]  Larry J. Eshelman,et al.  On Crossover as an Evolutionarily Viable Strategy , 1991, ICGA.

[11]  David E. Goldberg,et al.  The parameter-less genetic algorithm in practice , 2004, Inf. Sci..

[12]  Cláudio F. Lima,et al.  A review of adaptive population sizing schemes in genetic algorithms , 2005, GECCO '05.