Memetic Algorithm for Intense Local Search Methods Using Local Search Chains

This contribution presents a new memetic algorithm for continuous optimization problems, which is specially designed for applying intense local search methods. These local search methods make use of explicit strategy parameters to guide the search, and adapt these parameters with the purpose of producing more effective solutions. They may achieve accurate results, at the cost of requiring high intensity, making more difficult their application into a memetic algorithm. Our memetic algorithm approach assigns to each individual a local search intensity that depends on its features, by chaining different local search applications. With this technique of search chains, at each stage the local search operator may continue the operation of a previous invocation, starting from the final configuration reached by this one. The proposed memetic algorithm integrates the CMA-ES algorithm as their local search operator. We compare our proposal with other memetic algorithms and evolutionary algorithms for continuous optimization, showing that it presents a clear superiority over the compared algorithms.