A Hybrid Metaheuristic Based on Scaling Transformations
Applying metaheuristics to combinatorial optimization problems is a hot research field. To achieve a better optimal quality as well as a better optimal speed acceleration relies on intensification and diversification in metaheuristics. In this research, we propose a hybrid metaheuristic based on scaling transformations and demonstrate it by using the hybridization of the fitness component in evolutionary computation and the power transformation as an example. The main idea is to combine the requirements for intensification and diversification in different evolutionary stages with the concaveconvex properties of the power function to construct an adaptive evaluation function which gives the fitness value, i.e., an adaptive evaluation fitness function. The simulation results show great improvement on the maximum individual fitness value and average population fitness value, and therefore, the proposed approach is effective.
metaheuristics scaling transformations evolutionary computation evaluation function intensification and diversification
Weimin Peng Huifang Deng
School of Computer Science and Engineering,South China University of Technology,Guangzhou,China Scho School of Computer Science and Engineering,South China University of Technology,Guangzhou,China
国际会议
太原
英文
495-498
2011-02-26（万方平台首次上网日期，不代表论文的发表时间）