An Improved Multiobjective Quantum-Behaved Particle Swarm Optimization Based on Double Search Strategy and Circular Transposon Mechanism

Although multiobjective particle swarm optimization (MOPSO) has good performance in solving multiobjective optimization problems, how to obtain more accurate solutions as well as improve the distribution of the solutions set is still a challenge. In this paper, to improve the convergence performance of MOPSO, an improved multiobjective quantum-behaved particle swarm optimization based on double search strategy and circular transposon mechanism (MOQPSO-DSCT) is proposed. On one hand, to solve the problem of the dramatic diversity reduction of the solutions set in later iterations due to the single search pattern used in quantum-behaved particle swarm optimization (QPSO), the double search strategy is proposed in MOQPSO-DSCT. The particles mainly learn from their personal best position in earlier iterations and then the particles mainly learn from the global best position in later iterations to balance the exploration and exploitation ability of the swarm. Moreover, to alleviate the problem of the swarm converging to local minima during the local search, an improved attractor construction mechanism based on opposition-based learning is introduced to further search a better position locally as a new attractor for each particle. On the other hand, to improve the accuracy of the solutions set, the circular transposon mechanism is introduced into the external archive to improve the communication ability of the particles, which could guide the population toward the true Pareto front (PF). The proposed algorithm could generate a set of more accurate and well-distributed solutions compared to the traditional MOPSO. Finally, the experiments on a set of benchmark test functions have verified that the proposed algorithm has better convergence performance than some state-of-the-art multiobjective optimization algorithms.