A Novel Monarch Butterfly Optimization with Greedy Strategy and Self-Adaptive

Gai-Ge Wang,Xinchao Zhao,Suash Deb

DOI: https://doi.org/10.1109/ISCMI.2015.19

2015-01-01

Abstract:Recently, inspired by migration of monarch butterflies in the Northern American, a new kind of metaheuristic algorithm, called monarch butterfly optimization (MBO), is proposed for solving global optimization problems. It has been experimentally shown that MBO outperforms five state-of-the-art metaheuristic algorithms on most benchmarks. However, the main disadvantage of MBO is that it has poorer Std values and worse mean fitness on certain benchmarks. In this paper, in order to overcome this shortcoming, a greedy strategy is incorporated into the migration operation, and this incorporated strategy can only accept the monarch butterfly individuals that have better fitness than their parents. In addition, a self-adaptive crossover (SAC) operator is incorporated into the butterfly adjusting operator, and this SAC operator can significantly improve the diversity of population at later run phase of the search. In butterfly adjusting operator, the greedy strategy is also used to select the fitter monarch butterfly individual, which can accelerate convergent speed. Accordingly, a new version of Monarch Butterfly Optimization with Greedy strategy and self-adaptive Crossover operator (GCMBO) is proposed. Finally, the proposed GCMBO method is benchmarked by eighteen standard test functions. The results indicate that GCMBO method significantly outperforms the basic MBO method on almost all the test cases.