Memetic Algorithms with Extremal Optimization
Yong-Zai Lu, Yu‐Wang Chen, Min-Rong Chen, Peng Chen, Guo-Qiang Chen · Auerbach Publications eBooks · 2016
A particular class of global–LS hybrids named Memetic Algorithms (MA) are proposed, which are motivated by Dawkins’s theory. MAs are a class of stochastic heuristics for global optimization that combine the global-search nature of EA with LS to improve individual solutions. They have been successfully applied to hundreds of real-world problems such as optimization of combinatorial optimization, multiobjective optimization, and bioinformatics. The MAs combine the global-search ability of a random algorithm with the high-searching efficiency and precision of the LS algorithm, and realize the collaboration between different search methods, to achieve the unity of the global-search ability and LS efficiency. The structure of the hybrid EO–LM algorithm is based on standard EO, the characteristic of GS is added by propagating the individual solution with the LM algorithm during EO evolution. The proposed EO–LM solution has the abilities to avoid local minimum and perform detailed LS with both efficiency and robustness.