An Ethnic Group Evolution Algorithm Based on dual track coevolution

摘要

In ethnic group evolution algorithm (EGEA), the ethnic group clustering mechanism cannot only be used to analyze the structure of population but also be used to sift out valuable individuals from population. Based on this capability, we propose an improving EGEA with dual track coevolution mechanism consisting of an ethnic group evolution process and an experience learning process. The extra experience learning mechanism can discover experiential knowledge from typical individuals, which are gotten from each ethnic group, then use this kind of posterior knowledge to fix the structure of individuals so as to guide the evolution direction of population and improving the convergence speed of EGEA. The simulation tests of numerical optimization show the EGEA based on dual track coevolution mechanism is feasible and valid, which improve the searching efficiency of EGEA greatly.