MRMOGA: parallel evolutionary multiobjective optimization using multiple resolutions

Whereas multiobjective evolutionary algorithms have reached certain effectiveness in solving many real-world problems efficiency still remains as an open problem. One choice to reduce the execution time of the multiobjective evolutionary algorithms is their parallelization. This paper introduces a parallel MOEA which is based on the island model with heterogeneous nodes. This algorithm is characterized by encoding the solutions using a different resolution for each island. In this way, the search space is divided into well-defined overlapped regions in decision variable space.