A Dynamic Optimization Algorithm Based on Bi-Population Evolutionary Programming
Xiao Xiao · Jisuanji fangzhen · 2008
The optimal solution of dynamic optimization problem moves with time due to environment changing.In order to efficiently track the optimal solution,this paper proposes a dynamic optimization algorithm based on bi-population evolutionary programming.Local search population receives previous information and adopts Gaussian mutation,while the global search population,which is insulated from previous information,uses Cauchy mutation and transports its elitists to local search population.In the process of evolution,their population size is dynamically altered.The algorithm efficiently makes use of previous information,separates local search process from global search process,and is fit for solving the dynamic optimization problems where environment change mode is unclear.Experiments for three type dynamic optimization models show that the proposed method is more effective than the reinitializing strategy.