粒子群优化
计算机科学
算法
数学优化
多群优化
数学
算法设计
优化算法
元启发式
最优化问题
遗传算法
元优化
近似算法
进化算法
无导数优化
缩小
约束优化
进化计算
稳健性(进化)
信号处理
人工智能
作者
Tianzheng Wang,Jian Tang,Shengxiang Yang,Junfei Qiao
标识
DOI:10.1109/tevc.2026.3713919
摘要
Although extensive research has yielded numerous effective particle swarm optimization (PSO) variants through parameter tuning and structural modifications, these approaches frequently exhibit complex structures, reliance on prior knowledge, and limited adaptability. To address these limitations, a distance-aware and space-mapped PSO (DS-PSO) algorithm is proposed. The DS-PSO maintains structural simplicity while enhancing performance through improved search guidance and population diversity. Specifically, a distance-aware mechanism dynamically evaluates particle states based on their distances to the global best-found position obtained so far, adaptively tuning three control parameters to enhance local exploitation. Concurrently, a space-mapped mechanism introduces dimensional mutation to selected particles, thereby improving population diversity and avoiding premature convergence. Comprehensive experimental evaluations on the classical benchmark suite and the CEC 2022 Single-Objective Bound-Constrained Optimization (SO-BO) benchmark suite demonstrate that DS-PSO achieves better best-found solution quality and more stable performance. A practical engineering case is also employed to validate the effectiveness of the DS-PSO.
科研通智能强力驱动
Strongly Powered by AbleSci AI