论文部分内容阅读
快速追踪不同环境下pareto前沿,并保证最优解的收敛性和分布性,是动态多目标优化问题的关键.提出一种基于多领导粒子策略的动态多目标优化算法(Dynamic Multi-objective Particle Swarm Optimization Algorithm Based on Multiple Leaders Strategy,M LSDM PSO),在速度更新时,选择多个领导粒子来引导当前粒子飞行,寻找最优解;引入环境变化检测算子判断环境是否发生变化,并对环境变化做出响应.最后将该算法在三种不同类型的测试函数上进行测试,并将该算法与基于分解的动态多目标优化算法DMOEAD-M和基于多种群协同进化的动态多目标优化算法DCOEA进行比较,仿真结果表明,该算法有效的提高了对环境变化的追踪能力以及所得解的收敛性和分布性.
To quickly track the pareto frontier in different environments and ensure the convergence and distribution of the optimal solution is the key to the dynamic multi-objective optimization problem.This paper proposes a dynamic multi-objective Particle Swarm Optimization Algorithm Based on Multiple Leaders Strategy (MLSDM PSO). When the speed is updated, multiple leaders are selected to lead the current particle flight and find the optimal solution. The environment change detection operator is introduced to judge whether the environment changes or not, The algorithm is tested on three different types of test functions and the algorithm is compared with the dynamic multi-objective optimization algorithm DMOEAD-M based on decomposition and the dynamic multi-objective optimization algorithm DCOEA based on multi-population co-evolution The simulation results show that the proposed algorithm effectively improves the tracking ability of environmental changes and the convergence and distribution of the solutions.