用于电力行业决策支持的多AGENT技术研究

《用于电力行业决策支持的多AGENT技术研究》由会员上传分享，免费在线阅读，更多相关内容在学术论文-天天文库。

1、华东师范大学博士学位论文用于电力行业决策支持的多AGENT技术研究姓名：王勇申请学位级别：博士专业：系统分析与集成指导教师：黄国兴20070401摘要在分时电价的研究中，提出了基于Agent的分时电价优化模型，在总用电量不变条件下，达到了削峰填谷的目的。在电力负荷预测的研究中，提出了基于Agent的电力负荷预测算法，试验验证这种方法是有效的，并且采用多Agent的原理设计了一个电力负荷预测系统，实现了多种方法的负荷预测。本文的研究得到了上海市重点学科建设项目“电力企业信息化与决策支持”的资助，论文中的相关理论和技术已运用到该项目的

2、研究和开发中。关键词：人工智能，决策支持，多智能代理，电力ⅡABSTRACTDecisionSupportSystemsloss)isallinteractivesoftware．basedsystemintendedtohelpdecisionnlal(erscompileusefulinformationfromrawdatatoidentifyandsolveproblemsandmakedecisions．DSSbeginswithbuildiIlgmodel-drivenDSSinthelate1960s，theoryd

3、evelopmentsinthe1970s，andtheimplementationoffinancialplanningsystemsandgroupDSSintheearlyandmid80s．DatawarehoUSeS．0LAPandBusinessIntelligenceevolvedinthelate1980sandearly1990s．Nowdaysintelligencedecisionsupportsystem(msS)iSanewinformationsystemthatjoinsartificialintel

4、ligencetechnologyintodecisionsupportsystem．Agentistheindependenceintelligencesoftware，whichiSahottopicinthefieldofartificialintelligence．Amulti-agentsystemconsistsofanumberofagents，whichinteractwitheachotheronbehalfofeachagentwithdifferentgoalsandmotivations．Tosuccess

5、fullyinteract,也eywillrequiretheabilitytocooperate，coordinate，andnegotiatewitheachother．ElectricpowerdecisionsupportsystemiSwidelyusedinelectricpowermarkeLdemandsidemanagement(DSM)andelectric10adforecasting．AlthoughelectricpowerDSScallberealizedbyobjectorientedprogramm

6、ingtechnology,AgentOrientendProgramming(AOP)isn’t鹪incommonuseas00PI111eresearchonclassicaldecisionsupportsystemproblemfromAgentlogicstructuretoAgentinformationsystemiSdeficiency．Theresearchwillbehelpfulinelectricpowerdecisionsupportsystem．WeneedtousenewAgenttechnology

7、torealizetheIDSS，maybethetechnologywillbewidelyusedinadvanceddistributedelectricpowerDSSasOOPinthefuture．n壕dissertationfocusesonthetheoriesandmethodsaboutsoRwareintelligenceagent，multi—agentsystemstechnologyinintelligenceDSSandinelectricpowerDSS．n圮maincontributionsoft

8、hisdissertationareasfollows：Firstly，fuzzyreasonlogiconAgentispresented，bywhichpaperdeducedtheprocedureofagentgettingknowledg