一种基于多种群NSGA-Ⅱ的卷烟厂多AGV调度方法.pdfVIP

122一种基于多种群NSGA-II的卷烟厂多AGV调度方法

一种基于多种群NSGA-II的卷烟厂多AGV调度方法

NSGA-IIBasedonMulti-PopulationforMulti-AGVSchedulinginCigaretteFactory

李勇健吴珂雨江苏中烟工业有限责任公司南京卷烟厂袁江苏南京210019冤

摘要院在卷烟厂生产实践中袁自动导引车是助力生产自动化和智能化的重要运输工具遥针对AGV应用过程中出现的

AGV数量和系统作业效率的冲突袁以用车数量最小化和系统作业时间最小化为优化目标袁构建卷烟厂多AGV调度优化模

型遥为求解该问题袁设计一种多种群的NSGA-II算法袁引入多种交叉算子强化算法的全局探索能力袁构建三个群从不同方向

对Pareto最优解进行探索袁设计启发性邻域搜索策略提高算法收敛性遥基于实际工作环境袁构建了四个规模不同的测试算

例袁并通过与其他经典的多目标优化算法对比袁验证所提算法能够获取收敛性和分布性较好的Pareto解集袁为AGV调度提

供决策方案遥

关键词院卷烟厂曰多种群曰NSGA-II曰多AGV调度曰变邻域搜索

Abstract:AimingatthecontradictorybetweenthenumberofAGVsandtheoperatingefficiencyduringtheprocessof

AGVapplication,themulti-AGVschedulingoptimizationmodelfortheingredienttransportationsystemofacigarettefactoryis

established,tominimizethevehiclenumberandthemakesapn.Tosolvethisproblem,amulti-populationNSGA-IIalgorithm

isdesigned,whereavarietyofcrossoveroperatorsareintroducedtostrengthentheglobalexplorationabilityofthealgorithm,

threesubpopulationsareconstructedtoexploreParetooptimalsolutionsfromdifferentdirections,andtheheuristicneighbor鄄

hoodsearchstrategyisdesigned.Accordingtothepracticalworkingenvironment,fourtestinstanceswithdifferentscalesare

constructed.Bycomparingwithotherclassicalmulti-objectiveoptimizationalgorithms,itisverifiedthattheproposedalgorithm

canobtaintheParetosolutionsetwithgoodconvergenceanddistribution,providingdecisionschemesforAGVscheduling.

Keywords:cigarettefactory,multi-population,NSGA-II,multi-AGVscheduling,variableneighborhoodsearch

随着人们的健康意识不断提高袁吸烟人数和吸烟频率不断