Abstract: Aiming at the shortcomings of Marine Predator Algorithm (MPA),such as low optimization accuracy and slow convergence speed,an improved marine predator algorithm based on multi-strategy fusion was proposed.The initial population was updated by reverse differential variation before iteration.A two-population mechanism was proposed in the high-speed ratio stage.The global search was carried out based on the step size generated by Brownian motion and Weibull distribution respectively,and the dominant population renewal location was selected according to the fitness.Furthermore,the T-distribution adaptive disturbance strategy was introduced in the optimal individual determination stage,and the ocean memory was updated based on greedy selection.Based on 10 benchmark functions and part of CEC2017 functions,the performance of the proposed algorithm was evaluated by convergence analysis and Wilcoxon rank sum test.The experimental results showed that compared with the original marine predator algorithm,the optimization accuracy and convergence rate of the proposed algorithm were significantly improved in different dimensions,and it was significantly better than other comparison algorithms.Finally,the reliability and effectiveness of the proposed algorithm were further verified by two engineering constraint examples:pressure vessel design and vehicle side impact design.

Key words: meta heuristic algorithm, reverse differential variation, Weibull distribution, adaptive perturbation strategy, marine predator algorithm, pressure vessel design problem, car side impact design problem

摘要： 针对海洋捕食者算法寻优精度较低、收敛速度较慢等缺陷,提出一种多策略融合改进的海洋捕食者算法。首先,在迭代前通过反向差分变异对初始种群进行更新;其次,在高速比阶段提出一种双种群机制,分别以布朗运动和威布尔分布生成的步长进行全局搜索,并根据适应度大小选取优势种群更新位置;然后,在最优个体确定阶段引入t分布自适应扰动策略,同时基于贪婪选择更新海洋记忆。基于10个基准测试函数以及部分CEC2017函数,通过收敛性分析、Wilcoxon秩和检验进行性能评价,实验结果表明,相较于原始海洋捕食者算法,该算法在不同维度下的寻优精度与收敛速度均有明显改进,且显著优于其他对比算法。最后,通过压力容器设计与汽车侧面碰撞设计两个工程约束实例进一步验证了该算法的可靠性与有效性。

关键词: 元启发式算法, 反向差分变异, 威布尔分布, 自适应扰动策略, 海洋捕食者算法, 压力容器设计问题, 汽车侧面碰撞设计问题