Complex-Order Quantum-Behaved Particle Swarm Optimization With Double Jump-Out Strategy
Liping Chen, Xiaobo Wu, António Mendes Lopes, Da Wang, Panpan Gu, Min Zhu, YangQuan Chen · IEEE Transactions on Emerging Topics in Computational Intelligence · 2025
This paper presents a complex-order quantum-behaved particle swarm optimization (CoQPSO) algorithm, aiming at improving the local exploitation and global exploration abilities of existing optimization methods. The core of the CoQPSO is the adoption of complex-order derivatives within the particles' position adaptation mechanisms. The complex-order derivatives are computed with historical state information of the particles, being naturally suitable for the iterative procedure of intelligent optimization algorithms. To prevent convergence to local optima, a double jump-out strategy is designed so that particles can escape from local attractors. The influence of the algorithm's control parameters on its performance is assessed by means of sensitivity analysis, which comprises a number of value- and rank-based tests on a set of classical benchmark functions and CEC 2022 single-objective global optimization competition test suite. The CoQPSO is compared with other PSO variants in terms of the mean, standard deviation and best value of the solutions. Additionally, the Wilcoxon-rank sum test assesses algorithms' performance differences. The experimental results illustrate the superiority of the CoQPSO in finding optimal solutions.