Quasi opposite-based learning and double evolutionary QPSO with its application in optimization problems
Guang He,Xiao-li Lu
DOI: https://doi.org/10.1016/j.engappai.2023.106861
IF: 8
2023-08-03
Engineering Applications of Artificial Intelligence
Abstract:Although quantum-behaved particle swarm optimization (QPSO) algorithm has the advantages of few parameters and simple implementation, it suffers from these problems of low precision while calculating high-dimensional complex problems and being caught in local optimum during the later stage of iteration. To tackle these shortages effectively, a modified QPSO (QDQPSO) algorithm is presented. In order to improve the search efficiency and convergence speed of the algorithm, the idea of quasi opposite-based learning is used in the initialization stage. For enhancing the overall performance of the algorithm, the double evolutionary mechanism is applied to update the individual location during the iterative process. Furthermore, perturbation at global optimum position and bound constraint handling are considered to help the algorithm to escape from local optimum and maintain the diversity of population. According to the results obtained by QDQPSO and other nine optimization algorithms on 28 benchmark functions under different dimensions, it is found that QDQPSO performs better on the accuracy and stability of the optimal solution. Subsequently, Wilcoxon rank-sum test and Friedman test demonstrate significant advantages of the improved algorithm. Finally, QDQPSO algorithm displays superior performance in solving five practical optimization problems compared to several optimization methods.
automation & control systems,computer science, artificial intelligence,engineering, electrical & electronic, multidisciplinary