Scholarlink

<p>In particle swarm optimization (PSO), inertia weight plays a crucial role in balancing global and local search. However, fixed or simplistic adjustment strategies often struggle to effectively address complex optimization problems. To address this, an improved adaptive particle swarm optimization based on belief rule base (BRB-PSO) is proposed. Firstly, a universal BRB-based adaptive adjustment strategy for inertia weight is designed, enabling a dynamic balance between global and local search capabilities while ensuring adaptability to diverse optimization problems. Then, the evidential reasoning algorithm is integrated into the BRB to estimate the inertia weight increment and its current value accurately, which significantly improves the flexibility and generalization ability of rule inference. Furthermore, to prevent the algorithm from falling into local optima, the optimal values of each particle and the population are dynamically updated. Finally, the superiority of BRB-PSO is demonstrated by comparing the experimental results with other typical optimization algorithms, and its practical effectiveness is demonstrated in the parameter identification of the wind turbine pitch control system and ship steering machine.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

An Improved Adaptive Particle Swarm Optimization Based on Belief Rule Base

  • 期刊论文

  • 发表日期：

    2025年5月19日

查看全文
  • Xuelin Zhang,
  • Xiaobin Xu,
  • Song Sun,
  • Qinghua Zhong,
  • Hong He,
  • Georg Brunauer,
  • Felix Steyskal

摘要

In particle swarm optimization (PSO), inertia weight plays a crucial role in balancing global and local search. However, fixed or simplistic adjustment strategies often struggle to effectively address complex optimization problems. To address this, an improved adaptive particle swarm optimization based on belief rule base (BRB-PSO) is proposed. Firstly, a universal BRB-based adaptive adjustment strategy for inertia weight is designed, enabling a dynamic balance between global and local search capabilities while ensuring adaptability to diverse optimization problems. Then, the evidential reasoning algorithm is integrated into the BRB to estimate the inertia weight increment and its current value accurately, which significantly improves the flexibility and generalization ability of rule inference. Furthermore, to prevent the algorithm from falling into local optima, the optimal values of each particle and the population are dynamically updated. Finally, the superiority of BRB-PSO is demonstrated by comparing the experimental results with other typical optimization algorithms, and its practical effectiveness is demonstrated in the parameter identification of the wind turbine pitch control system and ship steering machine.