Improved Mind Evolutionary Algorithm Design Using Group Migration

Mind evolutionary algorithm (MEA) is a kind of evolutionary algorithm, simulating the human´s mind development. It has two operators, similartaxis and dissimilation. Swarm intelligence attempts to design algorithms or distributed problem-solving devices without concentrated control and global model. This paper is intrigued by the advantages of swarm intelligence, presents the mechanisms of the improved MEA and designs refreshing rule of information density and the group migration strategy according to share rule of social information, which delicately balance between good solution exploitation and new solution exploration, fastens convergence velocity and obtains global optimum. The performance on four different test functions is qualitatively analyzed. Experimental results show that the improved MEA with group migration based on swarm intelligence can rapidly converge at high quality solution with high stability and precision.