Solving multi-objective optimization problem using cuckoo search algorithm based on decomposition

In recent years, cuckoo search (CS) algorithm has been successfully applied in single-objective optimization problems. In addition, decomposition-based multi-objective evolutionary algorithms (MOEA/D) have high performance for multi-objective optimization problems (MOPs). Inspired by this, a new decomposition-based multi-objective CS algorithm is proposed in this paper. Two reproduction operators with different characteristics derived from the CS algorithm are constructed and they compose an operator pool. Then, a bandit-based adaptive operator selection method is used to determine the application of different operators. An angle-based selection strategy that achieves a better balance between convergence and diversity is adopted to preserve diversity. Compared with other improved strategies designed for MOEA/D on two suits of test instances, the proposed algorithm was demonstrated to be effective and competitive for MOPs.