多阶段改进的排球联赛算法及其在非线性优化问题中的应用  

A Multi-stage Improved Volleyball Premier League Algorithm and Its Application to Nonlinear Optimization Problems

作　　者：孙硕 马良[1] 刘勇[1] 尚春剑 SUN Shuo;MA Liang;LIU Yong;SHANG Chun-jian(School of Management,Shanghai University of Technology,Shanghai 200093,China)

机构地区：[1]上海理工大学管理学院,上海200093

出　　处：《系统工程》2024年第4期139-149,共11页Systems Engineering

基　　金：教育部人文社会科学研究青年基金项目(21YJC630087);上海市哲学社会科学规划课题(2019BGL014)。

摘　　要：针对排球联赛优化算法(VPL)在求解复杂优化问题时收敛效率低、容易陷入局部最优解等不足,提出一种改进的排球联赛算法(IIVPL)。在分析传统VPL算法存在的缺陷后,对算法种群初始解的生成和后续解的更新策略做了相应改进。首先,初始化阶段引入Logistic混沌映射策略,以增加种群多样性,扩大全局探索范围,提升求解效率;然后,在球队内更新解阶段引入自适应学习策略模拟球队根据自身水平向最优秀团队学习获取候选解,在队间更新阶段利用Chebyshev映射设计变异机制,以提高算法平衡全局搜索和局部开采的能力;最后,将IVPL算法与传统VPL算法、联赛冠军算法、足球联赛算法以及其他新型元启发式算法进行比较,利用20个标准测试函数进行仿真实验,采用MAE指标计算实验结果并排序。结果表明,IVPL算法在求解不同维度的单峰函数和多峰函数上具有比较优势。通过将IVPL算法成功应用于一阶倒立摆系统PID参数优化实例,验证了IVPL算法的有效性和实用性,体现了IVPL算法的实际应用价值。An improved volleyball premier league(IVPL)algorithm is proposed to address the shortcomings of the Volleyball premier league(VPL)algorithm in solving complex optimization problems with low convergence efficiency and easy to fall into local optimal solutions.After analyzing the defects of the original VPL algorithm,the generation of the initial solution of the algorithm population and the update strategy of the subsequent solutions are improved accordingly.First,the logistic chaotic mapping strategy is introduced in the initialization stage to increase the population diversity,expand the global exploration range and improve the efficiency of the search solution;then,an adaptive learning strategy is introduced in the intra-team update solution phase to simulate teams to learn from the best team to obtain candidate solutions according to their own level,and a variation mechanism is designed using Chebyshev mapping in the inter-team update phase to improve the algorithm's ability to balance global search and local exploitation capability;finally,IVPL algorithm is compared with original VPL algorithm,league championship algorithm,soccer league algorithm and other novel metaheuristic algorithms,and simulation experiments are conducted using 20 standard test functions,and the experimental results are calculated and ranked using MAE metrics.The results show that IVPL algorithm has comparative advantages in solving single-peaked functions and multi-peaked functions of different dimensions.The successful application of the IVPL algorithm to PID parameters optimization for a first-order inverted pendulum system demonstrates its effectiveness,practicality and practical value.

关 键 词：排球联赛优化算法 LOGISTIC混沌映射 自适应学习策略 Chebyshev变异机制 PID参数优化 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程]