Abstract.
The paper deals with a new hybrid search method that combines the precision of the local search methods with the global search capabilities of a Genetic Algorithm (GA). The strategy we propose borrows from the Lamarckian evolution paradigm with the restriction that local search is applied only to a fraction of the GA's population of individuals, under a global controlling mechanism that fights against redundant exploration of the search space. Experiments with our new strategy were conducted on a difficult optimization test set, in comparison with a performant variant of a GA. Results obtained show the superiority of our method both in terms of robustness and of the quality of solutions obtained.