An Improved Animal Migration Optimization Algorithm Based on Interactive Learning Behavior for High Dimensional Optimization Problem

摘要

Animal migration optimization(AMO) algorithm inspired by the behavior of animal migration is proposed recently. AMO shows good performance on the benchmark functions whose dimensionality is no more than 30. However, the performance of AMO is degraded rapidly when the dimensionality is larger than 30. In order to overcome this shortcoming, an improved animal migration algorithm (IAMO) based on interactive learning behavior is proposed in this paper. First, we introduce an interactive learning behavior that individuals will learn from each other by exchanging information. During the search process, the search step is dynamically adjusted. In this case, the intelligence of IAMO is higher than AMO. Second, a refined search method is used to search around the current solutions, and this method can enhance the search ability of the algorithm. Third, a birth-and-death mechanism is designed to avoid local optimum. The effectiveness of IAMO is verified on 100 dimensional benchmark functions, and the empirical results show that the performance of IAMO is promising.