| 摘要: |
| 当粒子群算法(PSO)解决多目标优化问题时，由于PSO有较快的收敛效果，使得种群在寻优过程中多样性不足，易使算法早熟收敛。为有效设计多目标粒子群算法,提出基于自适应网格混合机制的多目标粒子群算法(ammmMOPSO)。该算法采用自适应网格和混合机制的一种双重维护策略，以保证外部存档中的非劣解分布均匀，避免种群快速退化，影响粒子开发能力；利用混合机制中的加权策略在外部存档的非劣解中确定全局最优样本，增加了种群的多样性，提升粒子飞向真实Pareto前沿的概率；同时，为防止算法停滞，陷入局部最优的问题，还引入一个变异操作对粒子的位置进行动态变异，增强了粒子的探索能力。仿真实验结果表明：所提算法与其他3个国际经典的多目标粒子群算法相比,具有较优的收敛性和多样性,且有较好的空间化效果。 |
| 关键词: 多目标优化 粒子群算法 自适应网格 混合机制 变异操作 |
| DOI： |
| 分类号: |
| 基金项目: |
|
| Multi-objective Particle Swarm Optimization Based on Adaptive Mesh Mixing Mechanism |
|
ZOU Kang-ge,LIU Yan-min1,2
|
|
1.School of Mathematics and Statistics, Guizhou University, Guiyang 550025, China;2.2.School of Mathematics, Zunyi Normal College, Guizhou Zunyi 563006, China
|
| Abstract: |
| When particle swarm optimization (PSO) is used to solve multi objective optimization problems, the PSO has a fast convergence effect, which makes the diversity of population in the optimization process insufficient and makes the algorithm converge early. In order to effectively design multi objective particle swarm optimization algorithm, a multi objective particle swarm optimization (MOPSO) algorithm based on adaptive mesh mixing mechanism is proposed. The algorithm adopts a dual maintenance strategy of adaptive grid and mixing mechanism to ensure the uniform distribution of non inferior solutions in the external archive, and avoid rapid population degradation and affecting the development ability of particles. The weighted strategy in the mixing mechanism is used to determine the global optimal sample in the external non inferior solution, which increases the diversity of population and improves the probability of particles flying to the real Pareto frontier. At the same time, in order to prevent the algorithm from stagnating and falling into the local optimal problem, a mutation operation is introduced to dynamically change the position of particles, which enhances the exploration ability of particles. The simulation results show that the proposed algorithm has better convergence and diversity and better spatial effect than the other three classical multi objective particle swarm optimization algorithms. |
| Key words: multi objective optimization particle swarm optimization adaptive mesh mixed mechanism mutation operation |
