基于多目标粒子群优化的多功能车辆总线周期性扫描表的优化  

Optimization of the MVB Period Polling Table based on Multi-objective Particle Swarm Optimization

作　　者：陈佳凯[1] 韦巍[1] 

机构地区：[1]浙江大学电气工程学院,浙江杭州310027

出　　处：《铁道学报》2012年第11期60-66,共7页Journal of the China Railway Society

摘　　要：介绍多功能车辆总线(MVB)周期性扫描表的设计原则和优化方法。引入IEC-61375协议中定义的MVB周期性扫描表(PPT)相关基本概念,从合理利用总线、实时数据传输及突发数据应对的角度认为MVB周期性扫描表设计应遵循"均衡总体、异化个体"的设计原则。依照该原则,构造了周期性扫描表优化问题的数学模型,提出3个优化指标:最长周期相、最短周期相及异化度。然后针对性地设计一个多目标粒子群优化(MPSO)方法,该优化方法引入优化权重及效用函数协调、相互对立的优化目标,并提出粒子编码、外部种群更新、粒子记忆及位置更新、粒子变异的方法,最后经过实例说明该优化方法得到预期的优化结果。The design principle and optimization method of the Multifunction Vehicle Bus（MVB） Period Polling Table（PPT） were studied.The basic concepts relevant to MVB PPT defined in the IEC-61375 Protocol were introduced.The design of MVB PPT was set down to follow the principle of ＂overall balance,individual dissimilation＂ in consideration of rational utilization of the bus,real-time data transmission and burst data response.According to the principle,the mathematical model of PPT optimization was established.Three optimization indexes were given： the longest periodic phase,the shortest periodic phase and the degree of dissimilation.The multi-objective Particle Swarm Optimization（MPSO） method was specifically designed,which included the optimization weight and utility function to coordinate contrary optimization targets.The methods of particle coding,updating of external population,updating of particle memory and location and particle mutation were also indicated.Case-study proves the optimization method has reached expectation.

关 键 词：多目标粒子群优化 多功能车辆总线 周期性扫描表 

分 类 号：U285.5[交通运输工程—交通信息工程及控制]