A Multi-population Whale Optimization Algorithm Based on Orthogonal Learning

For solving the problems of low accuracy, low convergence speed and accidental premature of canonical whale optimization algorithm (WOA), a modified multi-population WOA based on orthogonal learning (MOL-WOA) is proposed. Firstly, the multi-population strategy is adopted. The diversity of the sub-populations can be improved by the generalized opposition-based learning (GOBL) strategy. Secondly, the nonlinear convergence factor and adaptive weight strategy are utilized to take account of the global search and local development. Finally, the orthogonal learning method is adopted to realize the communication among the sub-populations. At the same time, a generation jump operation is brought in for avoiding premature. Simulation experiments with 20 benchmark test functions, contrasted with the other two modified algorithms, MOL-WOA owns better performance on accuracy, convergence speed and ability to miss the local optimal.