Cloudde-based Distributed Differential Evolution for Solving Dynamic Optimization Problems

摘要

Although evolutionary algorithms (EAs) have been widely applied in static optimization problems (SOPs), it is still a great challenge for EAs to solve dynamic optimization problems (DOPs). This paper proposes a Cloudde-based differential evolution (CDDE) algorithm based on Message Passing Interface (MPI) technology to solve DOPs. During the evolutionary process, different populations are sent to different slave processes to perform mutation and crossover operations independently using different evolution strategies and then return to the master process to apply migration operation under an adaptive probability. Experimental studies were taken on several DOPs generated by the Generalized Dynamic Benchmark Generator (GDBG) which was used in 2009 IEEE Congress on Evolutionary Computation (CEC2009). The simulation result indicates that the proposed algorithm achieves promising performance in a statistical efficient manner.