We introduce in this article a new hybrid coevolutionary algorithm called hLCGA (hybrid Loosely Coupled Genetic Algorithm) that consists in combining a competitive coevolutionary genetic algorithm and a local search algorithm. We apply it to the Rosenbrock function optimization problem and compare the results of five hybrid variants to the original LCGA. We show the advantages of hybridizing a coevolutionary algorithm with local search algorithms in terms of solution quality and convergence speed.
Citation:
Gregoire Danoy, Pascal Bouvry, Tomy Martins, "hLCGA: A Hybrid Competitive Coevolutionary Genetic Algorithm," his, pp.48, Sixth International Conference on Hybrid Intelligent Systems (HIS'06), 2006