重庆邮电大学学报（自然科学版）2025，Vol.37Issue(1)：37-45,9.DOI:10.3979/j.issn.1673-825X.202312080407
基于改进差分进化算法的系统辨识方法
System identification method based on improved differential evolution algorithm
摘要
Abstract
For complex control systems,models obtained using traditional modeling methods often suffer from low accuracy and slow convergence.To address this issue,a system identification method based on an improved differential evolution(DE)algorithm is proposed.The method incorporates dynamic update strategies for mutation rates and crossover operators,real-time population updates,and optimized evolution termination conditions,significantly enhancing model accuracy and algorithm convergence speed.Compared to traditional least squares,second-order plus time-delay models,DE algorithms,and genetic algorithms,the proposed method improves accuracy by 94.91%,40.11%,23.33%,and 8.48%,respectively.Additionally,it achieves a nearly 3.2 times improvement in convergence speed over conventional genetic and DE algorithms.Experimental results demonstrate that the proposed method effectively mitigates the impact of interference signals and pro-duces precise system transfer function models.关键词
系统辨识/改进差分进化算法/超声换能器/传递函数模型Key words
system identification/improved differential evolution algorithm/ultrasonic transducer/transfer function model分类
信息技术与安全科学引用本文复制引用
吕证,陈博,李章勇,秦对..基于改进差分进化算法的系统辨识方法[J].重庆邮电大学学报（自然科学版）,2025,37(1):37-45,9.基金项目
国家自然科学基金项目(11904042) （11904042）
中国博士后科学基金项目(2022MD723728) （2022MD723728）
重庆市自然科学基金项目(CSTB2023NSCQ-MSX0861)National Natural Science Foundation of China(11904042) （CSTB2023NSCQ-MSX0861）
Postdoctoral Science Foundation of China(2022MD723728) （2022MD723728）
Natural Science Foundation Project of Chongqing(CSTB2023NSCQ-MSX0861) （CSTB2023NSCQ-MSX0861）
