基于Agent协作机制的过程系统控制研究
摘要
随着工业生产规模的扩大和生产工艺技术的改进,所需控制的对象越来越复杂,常伴有时变和不确定性,难以用已有的数学解析方法精确描述被控对象特性;另一方面,控制性能要求却很高,控制目标呈多层次、多样化的特点。这一切都使传统控制理论遇到了严重挑战,必须发展新的控制理论。沿着控制理论和技术的发展历史可以看出,其产生、存在和发展的条件必然与当时的物质和技术基础相适应。计算机技术、人工智能技术和信息论的迅速发展,为控制理论和技术进行变革性发展提供了条件。本文正是在这样的背景下,针对复杂的过程对象,着重研究基于Agent的智能控制系统的实现方法。
本文以Agent协作机制为出发点,将结果共享的协作策略应用于关联过程系统中。运用Agent理论将系统划分为若干Agent,一方面每个Agent独立控制自身回路,另一方面系统的各Agent之间采用结果共享的协作策略,获得系统的局部和全局信息,通过控制系统仿真得到了良好的控制效果。最后在实验室DCS、PLC装置平台下的实例研究,结果表明,基于Agent的控制结构有效地解决了关联对象控制系统中不同回路间的解耦和协调问题,是实现复杂过程系统控制的一种可行方案。
另外,本文还把研究多人决策问题的博弈理论引入到多Agent之间的协作行为中,提出了基于不完全信息静态博弈的多Agent之间的协作模型,并对协作算法的实现进行了探讨。
本文的研究面向生产需求,探讨将人工智能控制领域的先进理论应用于工业过程控制的途径,研究工作为丰富复杂工业过程系统的控制方法提供了有价值的结果。
Along with industrial production scale expansion and production processing technology improvement,the control object is become more and more complex, time-variable and the uncertainty.It is difficulty for the mathematical analysis method which already has to description object performance accurately.On the other hand,the control performance requirement is very high,The control goal becomes diverse and multi-level.All these caused the traditional control theory to face the serious challenge and have to develop the new control theory.In history, Control technology and theory comes into being,exist and develop must be adapted with the material and technology which has been.Developing rapidly in the computer technology,the artificial intelligence technology and the information theory,it provides for the transformation development in the control technology and theory.In this paper,to a kind of complex process control system,an agent based intelligence control system is put forward:
According to Agent cooperation mechanism,result-sharing cooperation collaboration strategy is applied in associated process control system.Using the Agent theory,the system is divided into several agents.On the one hand agents control own loop independently,on the other hand each agent makes use of the result sharing cooperation strategy to obtain local and global information from the system,with satisfying control results obtained.Finally,the proposed agent based cooperation mechanism control strategy is executed in process system which consists of DCS,PLC.The result show that the problem of decouple and coordination in the different loop of the complex process control system has been solved effectively by utilizing the structure of Agent.
In-addition,the paper researched cooperation and coordination of MAS by game theory and its realization also has been mentioned.
The essay combines the theory research and practice together.Brought artificial intelligence control theory from other related field into industry process control field,the paper pointed out a new way for the research of automatic control theory applied in industry complex process system,which has provided the valuable result.
本文以Agent协作机制为出发点,将结果共享的协作策略应用于关联过程系统中。运用Agent理论将系统划分为若干Agent,一方面每个Agent独立控制自身回路,另一方面系统的各Agent之间采用结果共享的协作策略,获得系统的局部和全局信息,通过控制系统仿真得到了良好的控制效果。最后在实验室DCS、PLC装置平台下的实例研究,结果表明,基于Agent的控制结构有效地解决了关联对象控制系统中不同回路间的解耦和协调问题,是实现复杂过程系统控制的一种可行方案。
另外,本文还把研究多人决策问题的博弈理论引入到多Agent之间的协作行为中,提出了基于不完全信息静态博弈的多Agent之间的协作模型,并对协作算法的实现进行了探讨。
本文的研究面向生产需求,探讨将人工智能控制领域的先进理论应用于工业过程控制的途径,研究工作为丰富复杂工业过程系统的控制方法提供了有价值的结果。
Along with industrial production scale expansion and production processing technology improvement,the control object is become more and more complex, time-variable and the uncertainty.It is difficulty for the mathematical analysis method which already has to description object performance accurately.On the other hand,the control performance requirement is very high,The control goal becomes diverse and multi-level.All these caused the traditional control theory to face the serious challenge and have to develop the new control theory.In history, Control technology and theory comes into being,exist and develop must be adapted with the material and technology which has been.Developing rapidly in the computer technology,the artificial intelligence technology and the information theory,it provides for the transformation development in the control technology and theory.In this paper,to a kind of complex process control system,an agent based intelligence control system is put forward:
According to Agent cooperation mechanism,result-sharing cooperation collaboration strategy is applied in associated process control system.Using the Agent theory,the system is divided into several agents.On the one hand agents control own loop independently,on the other hand each agent makes use of the result sharing cooperation strategy to obtain local and global information from the system,with satisfying control results obtained.Finally,the proposed agent based cooperation mechanism control strategy is executed in process system which consists of DCS,PLC.The result show that the problem of decouple and coordination in the different loop of the complex process control system has been solved effectively by utilizing the structure of Agent.
In-addition,the paper researched cooperation and coordination of MAS by game theory and its realization also has been mentioned.
The essay combines the theory research and practice together.Brought artificial intelligence control theory from other related field into industry process control field,the paper pointed out a new way for the research of automatic control theory applied in industry complex process system,which has provided the valuable result.
引文
[1]李飒.基于移动Agent的分布式煤矿安全信息管理系统[J].煤矿安全,2007,38(7):90-92.
[2]范瑾,诸叶平.基于Agent技术的玉米生产管理协同决策系统研究[J].中国农学通报,2007,23(7):572-576.
[3]杨世杰,蒋国瑞,黄梯云.一种基于Multi-Agent的供应商选择系统[J].计算机应用与软件,2007,24(7):114-115.
[4]朱艳辉,阳爱民,王平.基于Agent的中文信息智能检索系统[J].计算机应用与软件,2007,24(5):42-44.
[5]黄伟雄,王晓东.基于Agent的智能交通信号控制系统研究[J].福建电脑,2007(1):5-6.
[6]秦斌,吴敏,王欣,阳春华.基于MAS的分布式焦炉集气管压力解耦控制[J].控制理论与应用,2006,23(6):961-966.
[7]张洁,高亮,李培根.Multi-Agent技术在先进制造中的应用[M].北京:科学出版社,2004:10-12;170-171.
[8]Rao A S,Georgeff M P.Modeling Rational Agents within a BDI-architecture[C].In Proceedings of the second International Conference on Principles of Knowledge Representation and Reasoning.1991:473-484.
[9]马光伟,徐晋晖.Agent思维状态模型[J].软件学报,1999,10(4):342-348.
[10]窦立谦.多Agent协作与协调的研究[D].天津:天津大学硕士学位论文,2005.
[11]Jennings N R.Commitment,Conventions:the Foundation of Coordination in Multi-Agent System[J].The Knowledge Engineering Preview,1993.
[12]Grose B.Collaborative System[M].AI magazine,1995,17(2):67-85.
[13]Brooks,R.Elephants Don't Play Chess[M].Robotics and Autonomous Systems,1990(6):3-15.
[14]F Ho,M Kamel.Learning Coordination Strategies for Cooperative Multiagent Systems [J].Machine Learning,1998,33(2-3):155-177.
[15]Garland A,Alterman R.Autonomous agents that learn to better coordinate[J].Autonomous Agents and Multi-Agent Systems,2004,8(3):267-301.
[16]罗健,王成良.基于KQML的Agent通信体系结构设计[J].重庆大学学报,自然科学 版,2004,27(3):56-60.
[17]张德同,周明全,耿国华,郑霞.FIPA Agent通讯规范分析[J].计算机科学,2004,31(7):166-168.
[18]徐明,庄毅.基于XML的MAS的研究与实现[J].计算机科学,2006,33(2):205-207.
[19]高国军,段永强,张申生.基于CORBA和多代理技术的可重构企业信息系统[J].计算机集成制造系统,2000,6(3):25-29.
[20]Swaminathan J M,Smith S F,Sadeh N M.A Multi-Agent framework for supply chain dynamics[J].In:Proceedings of Manufacturing Systems,1988,7(2):83-94.
[21]Shen W,Norrie D H.A Hybrid Agent-Oriented Infrastructure for Modeling Manufacturing Enterprises[EB/OL].http://www.ksi.cpsc.ucalgary.ca.
[22]廖强,周凯,张伯鹏.基于现场总线的Multi-Agent作业车间动态调度问题的研究[J].中国机械工程,2000,22(7):757-759.
[23]刘金琨,王树青.钢铁工业过程多智能体系统研究[J].信息与决策,1998,27(6):427-432.
[24]刘金琨,王树青.复杂实时动态环境下的多智能体系统[J].控制与决策,1998,7,13(增刊):385-390.
[25]张宏伟,吴爱国,盛涛.多智能体分布式智能控制在VAV空调系统中的应用[J].中国工程科学,2006,8(7):58-62.
[26]Lane D M.Distributed problem solving and real time mechanisms in robot architectures[J].Engineering Application Intelligence,1994,7(2):105-117.
[27]陈仁际,王韬等.分布式多机器人装配系统任务合作规划算法研究[J].中国机械工程,2000,11(4):418-421.
[28]刘大有,杨鲲,陈建中.Agent研究现状与发展趋势[J].软件学报,2000,11(3):315-321.
[29]Hong Liu,Ming xi Tang,John Hamilton Frazer.Supporting evolution in a multi-agent cooperative design environment supporting evolution in a multi-agent cooperative design environment[J].Advances in Engineering Software 2002(33):319-329.
[30]石慧,徐从富,刘勇,董金.Agent通信语言KQML的实现及应用[J].计算机工程与应用,2005,41(13):94-97.
[31]Wang Zhi-he,Yu Mei,Yin Xu-ri.A collaboration mechanism on positive interactions in AODE[J].Journal of Northwest Normal University(Natural Science),1999,35(1):27-31.
[32]HE Yan xiang,DU zhuo min,LIU Chao yang et al.A strategy for Multi-Agent Collaboration in MADCE[J].J of Wuhan Univ(Nat Sci Ed),2001,47:42-46.
[33]陈玲.基于Multi-Agent的复杂工业过程控制系统及其在PROFIBUS上实现的研究[D].上海:东华大学硕士学位论文,2004.
[34]王化冰,张柳芳.过程控制系统及其策略[J].煤矿机械,2005(5):3-5.
[35]裴建兵,商超平.DCS及其发展趋势[J].数字化工,2005(8):4-5.
[36]王锦标.浅议石化企业DCS、FCS的应用和发展[J].数字石油和化工,2006(11):2-5.
[37]唐瑛.现代工业过程控制系统的技术分析[J].科技信息,学术版,2007(17):84-84.
[38]朱建明.PROFIBUS在过程控制系统中的应用[J].中国教育研究与创新,2007,4(4):32-33.
[39]陶明璋.计算机集成过程系统中的智能控制[J].电子计算机外部设备,1996,20(5):48-50.
[40]冯纯伯.复杂工业过程的控制问题[J].控制理论与应用.1992,9(3):312-319.
[41]Hewitt C.A universal modular actor formalism for AI[C].In Proceedings of the 3~(rd)International Joint Conference on Artificial Intelligence(IJCAI-73),1973:235-245.
[42]Wooldridge M,Jennings N.Intelligent Agents:theory and practice[J].The Knowledge Engineering Review,1995,10(2):115-152.
[43]Shoham Y.Agent oriented programming[J].Artificial Intelligence,1993,(60):51-92.
[44]王荣波,周昌乐.移动Agent研究综述[J].计算机应用研究,2001,(6):9-11.
[45]韩晓征.基于分布式控制系统的MAS中若干问题研究[J].合肥:合肥工业大学硕士论文,2002.
[46]蔡自兴,周翔,李枚毅.基于功能/行为集成的自主式移动机器人进化控制体系结构[J].机器人,2000,22(3):169-175.
[47]叶媛媛,薛宏涛,沈林成.交互式智能体体系结构应用研究[J].计算机与现代化,2001,(1):27-31.
[48]董红斌,孙羽.多Agent系统的现状与进展[J].计算机应用研究,2001,18(1):54-56.
[49]Durfee E H,Lesser V R,Corkill D D.Trends in cooperative distributed problem solving network.Proc[J].IEEE Transudations Knowledge and Data Engineering,1989,KOE-11(1):63-83.
[50]Jennings N R,Sycara K,Wooldridge M.A roadmap of Agent research and development[J]. Autonomous Agent sand Multi-Agent Systems,1998,10(2):199-215.
[51]史忠植.智能主体及其应用[M].科学出版社,2000.
[52]Simth R G.The contract net protocol:High-level communication and control in a distributed problem solver[J].IEEE Transaction on Computer,1980,No,12.
[53]Durfee E H,Lesser V R.Using partial global plans to coordinated distributed problem solver[C].Processing of the 10th international Joint Conference on Artificial Intelligence (IJ-CAI-87),Milan,Italy,1987:875-883.
[54]刘向军,刘世平,张洁等.MAS通信与协作机制构造[J].机械设计与制造工程,2002.
[55]Lesser V R,Erman L D.Distributed interpretation:a model and experiment[J].IEEE Transaction on Computers,1981,29:1144-1163.
[56]张维迎.博弈论与信息经济学[M].上海:上海人民出版社,1996.
[57]肖条军.博弈论及其应用[M].上海:上海三联出版社,2004.
[58]袁爱进,曹立明,王小平.一种基于博弈论的多Agent交互模型[J].计算机工程,2005,31(6):50-51.
[59]魏武,江岸,谢赛,郭艳玲.基于遗传再励和博弈论的多Agent交通控制系统优化和协调[J].长沙理工大学学报:自然科学版,2007,4(1):24-28.
[60]王俊普,徐杨,陈逼,李滔.基于Agent的智能控制系统的实现方法研究[J].自动化与仪表,2001,16(3):26-29.
[61]中国工控网[EB/OL],Available:http://www.gongkong.com/tech/.
[62]蒋蔚孙,俞金寿.过程控制工程[M].北京:科学出版社,1998:1-5.
[2]范瑾,诸叶平.基于Agent技术的玉米生产管理协同决策系统研究[J].中国农学通报,2007,23(7):572-576.
[3]杨世杰,蒋国瑞,黄梯云.一种基于Multi-Agent的供应商选择系统[J].计算机应用与软件,2007,24(7):114-115.
[4]朱艳辉,阳爱民,王平.基于Agent的中文信息智能检索系统[J].计算机应用与软件,2007,24(5):42-44.
[5]黄伟雄,王晓东.基于Agent的智能交通信号控制系统研究[J].福建电脑,2007(1):5-6.
[6]秦斌,吴敏,王欣,阳春华.基于MAS的分布式焦炉集气管压力解耦控制[J].控制理论与应用,2006,23(6):961-966.
[7]张洁,高亮,李培根.Multi-Agent技术在先进制造中的应用[M].北京:科学出版社,2004:10-12;170-171.
[8]Rao A S,Georgeff M P.Modeling Rational Agents within a BDI-architecture[C].In Proceedings of the second International Conference on Principles of Knowledge Representation and Reasoning.1991:473-484.
[9]马光伟,徐晋晖.Agent思维状态模型[J].软件学报,1999,10(4):342-348.
[10]窦立谦.多Agent协作与协调的研究[D].天津:天津大学硕士学位论文,2005.
[11]Jennings N R.Commitment,Conventions:the Foundation of Coordination in Multi-Agent System[J].The Knowledge Engineering Preview,1993.
[12]Grose B.Collaborative System[M].AI magazine,1995,17(2):67-85.
[13]Brooks,R.Elephants Don't Play Chess[M].Robotics and Autonomous Systems,1990(6):3-15.
[14]F Ho,M Kamel.Learning Coordination Strategies for Cooperative Multiagent Systems [J].Machine Learning,1998,33(2-3):155-177.
[15]Garland A,Alterman R.Autonomous agents that learn to better coordinate[J].Autonomous Agents and Multi-Agent Systems,2004,8(3):267-301.
[16]罗健,王成良.基于KQML的Agent通信体系结构设计[J].重庆大学学报,自然科学 版,2004,27(3):56-60.
[17]张德同,周明全,耿国华,郑霞.FIPA Agent通讯规范分析[J].计算机科学,2004,31(7):166-168.
[18]徐明,庄毅.基于XML的MAS的研究与实现[J].计算机科学,2006,33(2):205-207.
[19]高国军,段永强,张申生.基于CORBA和多代理技术的可重构企业信息系统[J].计算机集成制造系统,2000,6(3):25-29.
[20]Swaminathan J M,Smith S F,Sadeh N M.A Multi-Agent framework for supply chain dynamics[J].In:Proceedings of Manufacturing Systems,1988,7(2):83-94.
[21]Shen W,Norrie D H.A Hybrid Agent-Oriented Infrastructure for Modeling Manufacturing Enterprises[EB/OL].http://www.ksi.cpsc.ucalgary.ca.
[22]廖强,周凯,张伯鹏.基于现场总线的Multi-Agent作业车间动态调度问题的研究[J].中国机械工程,2000,22(7):757-759.
[23]刘金琨,王树青.钢铁工业过程多智能体系统研究[J].信息与决策,1998,27(6):427-432.
[24]刘金琨,王树青.复杂实时动态环境下的多智能体系统[J].控制与决策,1998,7,13(增刊):385-390.
[25]张宏伟,吴爱国,盛涛.多智能体分布式智能控制在VAV空调系统中的应用[J].中国工程科学,2006,8(7):58-62.
[26]Lane D M.Distributed problem solving and real time mechanisms in robot architectures[J].Engineering Application Intelligence,1994,7(2):105-117.
[27]陈仁际,王韬等.分布式多机器人装配系统任务合作规划算法研究[J].中国机械工程,2000,11(4):418-421.
[28]刘大有,杨鲲,陈建中.Agent研究现状与发展趋势[J].软件学报,2000,11(3):315-321.
[29]Hong Liu,Ming xi Tang,John Hamilton Frazer.Supporting evolution in a multi-agent cooperative design environment supporting evolution in a multi-agent cooperative design environment[J].Advances in Engineering Software 2002(33):319-329.
[30]石慧,徐从富,刘勇,董金.Agent通信语言KQML的实现及应用[J].计算机工程与应用,2005,41(13):94-97.
[31]Wang Zhi-he,Yu Mei,Yin Xu-ri.A collaboration mechanism on positive interactions in AODE[J].Journal of Northwest Normal University(Natural Science),1999,35(1):27-31.
[32]HE Yan xiang,DU zhuo min,LIU Chao yang et al.A strategy for Multi-Agent Collaboration in MADCE[J].J of Wuhan Univ(Nat Sci Ed),2001,47:42-46.
[33]陈玲.基于Multi-Agent的复杂工业过程控制系统及其在PROFIBUS上实现的研究[D].上海:东华大学硕士学位论文,2004.
[34]王化冰,张柳芳.过程控制系统及其策略[J].煤矿机械,2005(5):3-5.
[35]裴建兵,商超平.DCS及其发展趋势[J].数字化工,2005(8):4-5.
[36]王锦标.浅议石化企业DCS、FCS的应用和发展[J].数字石油和化工,2006(11):2-5.
[37]唐瑛.现代工业过程控制系统的技术分析[J].科技信息,学术版,2007(17):84-84.
[38]朱建明.PROFIBUS在过程控制系统中的应用[J].中国教育研究与创新,2007,4(4):32-33.
[39]陶明璋.计算机集成过程系统中的智能控制[J].电子计算机外部设备,1996,20(5):48-50.
[40]冯纯伯.复杂工业过程的控制问题[J].控制理论与应用.1992,9(3):312-319.
[41]Hewitt C.A universal modular actor formalism for AI[C].In Proceedings of the 3~(rd)International Joint Conference on Artificial Intelligence(IJCAI-73),1973:235-245.
[42]Wooldridge M,Jennings N.Intelligent Agents:theory and practice[J].The Knowledge Engineering Review,1995,10(2):115-152.
[43]Shoham Y.Agent oriented programming[J].Artificial Intelligence,1993,(60):51-92.
[44]王荣波,周昌乐.移动Agent研究综述[J].计算机应用研究,2001,(6):9-11.
[45]韩晓征.基于分布式控制系统的MAS中若干问题研究[J].合肥:合肥工业大学硕士论文,2002.
[46]蔡自兴,周翔,李枚毅.基于功能/行为集成的自主式移动机器人进化控制体系结构[J].机器人,2000,22(3):169-175.
[47]叶媛媛,薛宏涛,沈林成.交互式智能体体系结构应用研究[J].计算机与现代化,2001,(1):27-31.
[48]董红斌,孙羽.多Agent系统的现状与进展[J].计算机应用研究,2001,18(1):54-56.
[49]Durfee E H,Lesser V R,Corkill D D.Trends in cooperative distributed problem solving network.Proc[J].IEEE Transudations Knowledge and Data Engineering,1989,KOE-11(1):63-83.
[50]Jennings N R,Sycara K,Wooldridge M.A roadmap of Agent research and development[J]. Autonomous Agent sand Multi-Agent Systems,1998,10(2):199-215.
[51]史忠植.智能主体及其应用[M].科学出版社,2000.
[52]Simth R G.The contract net protocol:High-level communication and control in a distributed problem solver[J].IEEE Transaction on Computer,1980,No,12.
[53]Durfee E H,Lesser V R.Using partial global plans to coordinated distributed problem solver[C].Processing of the 10th international Joint Conference on Artificial Intelligence (IJ-CAI-87),Milan,Italy,1987:875-883.
[54]刘向军,刘世平,张洁等.MAS通信与协作机制构造[J].机械设计与制造工程,2002.
[55]Lesser V R,Erman L D.Distributed interpretation:a model and experiment[J].IEEE Transaction on Computers,1981,29:1144-1163.
[56]张维迎.博弈论与信息经济学[M].上海:上海人民出版社,1996.
[57]肖条军.博弈论及其应用[M].上海:上海三联出版社,2004.
[58]袁爱进,曹立明,王小平.一种基于博弈论的多Agent交互模型[J].计算机工程,2005,31(6):50-51.
[59]魏武,江岸,谢赛,郭艳玲.基于遗传再励和博弈论的多Agent交通控制系统优化和协调[J].长沙理工大学学报:自然科学版,2007,4(1):24-28.
[60]王俊普,徐杨,陈逼,李滔.基于Agent的智能控制系统的实现方法研究[J].自动化与仪表,2001,16(3):26-29.
[61]中国工控网[EB/OL],Available:http://www.gongkong.com/tech/.
[62]蒋蔚孙,俞金寿.过程控制工程[M].北京:科学出版社,1998:1-5.