|
Descripción
|
|
|---|---|
| Forthoserailwaystationswithoutbeingautomated,railwaytrafficdispatchingstilldependsondispatchers,especiallyunderdisturbedcircumstances.In this study, anagent-basedsupport system, namedDAgent,is developedto assisthumandispatchersto makedecisionsin stationoperation.Tothis end,the commonknowledgeandpossibledifficultiesconcerningastationdispatcherin his/herroutine workare firstlystudied, andthe D-Agentis proposedwiththe purposeofworkingoutpracticablesolutions to these challengingtasks asadispatcherdoes.Thenthe generalmodelofthe D-Agentis established,containingfivebasicmodules:local database,knowledgebase,skillbase,reasoningmechanismandcommunicationinterfaces. Theinternal skillsofthe D-Agentaredesignedto executevarioustasks in different scenarios. Besides,askill extensionofthe D-Agentwithmathematicalformulations is particularlydiscussedin this paper,to findfeasible andoptimaltraffic controlsolutionsin disturbancesituationssuch astrain delaysandroute conflicts.TheD-Agentis designedto learn from its ownexperimentalhistory in applyingdifferentskills,andevaluatetheskillsbypreferenceweightsofalternativesolutionsin aparticulartask. Thisprocedureallowstheagentto havepotentialfor continuousimprovement. Toverifythe applicabilityoftheproposedsupportsystem,aD-Agentfor aterminalstationofsubwayis simulated.The numericalexampleoftraindelaysandrouteconflictsshowsthattheD-Agentcangenerallyperformasa stationdispatcherin fulfilling thespecifictasks,estimatethetrafficstatein differentoperationstrategies andsupport the decision-makingoffavored solutions. Significantly,it indicates that the mathematical methodscanalsobeenemployedbyanintelligent agent. | |
|
Internacional
|
Si |
|
JCR del ISI
|
Si |
|
Título de la revista
|
Expert Systems With Applications |
|
ISSN
|
0957-4174 |
|
Factor de impacto JCR
|
2,981 |
|
Información de impacto
|
Datos JCR del año 2015 |
|
Volumen
|
|
|
DOI
|
|
|
Número de revista
|
|
|
Desde la página
|
39 |
|
Hasta la página
|
52 |
|
Mes
|
SIN MES |
|
Ranking
|
|