中国科学院软件研究所机构知识库
Advanced  
ISCAS OpenIR  > 中科院软件所  > 中科院软件所
题名:
Agent系统通讯理论及组织结构的研究
作者: 王栩
答辩日期: 2000
专业: 计算机软件理论
授予单位: 中国科学院软件研究所
授予地点: 中国科学院软件研究所
学位: 博士
关键词: Agent通讯语言 ; 会话模型 ; 会话协议 ; 协议验证 ; 软件工程
摘要: Agent是计算机中活动的软件实体。它们置身于特定的计算环境中,各自控制一些资源,负责完成不同的任务。单个Agent已经能够完成一些较有意义的工作;但是,更有价值、研究前景更为广阔的却是由一群自主协调、相互配合Agent个体组成的多Agent系统。尤其考虑到未来的 Agent计算环境将是一种开放、分布、自治的网络计算环境。当前,多Agent系统中进行协调和配合的主要手段是采用Speech-act语言的相互通讯。Speech-act通讯在多Agent系统中所起的作用就如话语交流在人类群体中所起的作用。本论文正是抓住Speech-act通讯这个关键对多Agent系统进行研究。我们的研究将专注于Agent的通讯特性;因此,我们称被研究的Agent系统为Agent通讯系统。具体地,本论文对Agent通讯系统在以政几方面进行了研究:1. 针对用为通讯基础的Speech-act语言,提出了一种以会观察者心灵状态为核心的语义方法。采用这一方法,为一简单的样例语言SAIL,给出了操作语义,并说明了使用此语义对Agent系统中通讯行为进行语义恰当性检验的方法。2. 针对Agent间的动态通讯行为,指出,如同人类的会话行为要遵守一些隐含的结构性规律一样,Agent在通讯时也需遵循某些结构化准则。这些准则体现为:主动权交替控制,多话题嵌套进行,话题中隐含协议等等。这些准则构成了我们关于Agent通讯的ARC模型。3. 针对多Agent系统中Agent间的组织关系,提出了以机构这种可伸缩的组织单位为核心的组织域理论。并且我们还进一步引入了“内窥“机制,使得我们可以突破等级式的组织形式,采用更复杂、更灵活的组织形式。4. 基于ARC模型的组织域理论,设计了一种形式化的Agent通讯系统规范描述语言ASML,此语方将Agent通讯系统中静态组织结构的描述和动态通讯行为的描述较好地结全在了一起。5. 针对Agent通讯系统(尤其是动态通讯行为部分)的分析和设计中面临的挑战,我们借鉴会话分析理论,提出了从会话场景中解析会话,话题和话题协议的方法。6. 提出了基于测试机的方法来验证会话协议的实现策略,并证明了测试机定理。在总体上,本论文是以一种软件工程的方式对Agent通讯系统进行研究。最终,我们将以上提出的各种语言、模型和方法应用于分析的设计一个实例系统,并对结果过行了评价,把我们的方法与传统的和其他Agent的软件工程方法进行了比较,指出其优缺点及未来工作的方向。
英文摘要: Agent are active computational entities situated in complex environments, where they utilize available resources to undertake assigned tasks. In simple cases, individual agents are capable enough to implement very meaningful applications. However, to fully exploit agents' potentiality (especially in a distributed, autonomous, and open environment), we need group of agents which could coordiante their behaviors and be functioning together. Almost all the coordination and cooperation activities amongst agents involve, in one way or another, the agents' ability to communicate in speech-act languages. Speech-act communication is playing the role amongst agents what human speech is playing amongst a group of people. This thesis is a study of multi-agent system from the very perspective of speech-act communication. As the communicative aspect of agents is the focus of our work, the other aspects of agents are necessarily treated as of lesser importance. So, the multi-agent systems under study by us are also called agent communication systems. As a broad outline, we can say our work on agent communication systems is carried on in the following five respects. 1. We introduce a semantics framework based on conversational observer's mental state four understanding and designing speech-act languages, and have used the framework to give an operational semantics for a light-weight exemplar speech-act language, SAIL. 2. We make it clear that, just as there are some inherent structural principle underlying human conversation, agent communication should as well abide by some structurzation rules. Among these rules are, the turns-taking system of a conversation, the nested occurrences of topics, and the protocols regulating communicative behaviors within a topic. All these and some other things around conversation collectively comprise what we called ARC Model for agent communication. 3. Agent systems (esp. large-scale multi-agent systems) usually have very complex organizational relationships within. To structure these relationships, we propose a flexible and scalable construct, scope, as the basic organizational unit for agent communication systems. Scope can not only organize a collection of roles around an agent, it also organize a group of agents within an organization, which may function just like an agent by participating in conversations. Scopes, organizations, agents and roles collectively comprise what we called Organizational Scope Theory. 4. To support the development of agent communication systems based on ARC Model and Organizational Scope Theory, we need a modeling language tailored specifically for the specification of agent communication system. As a result, in this thesis, we have designed the ASML language to fulfill this requirement. 5. The development of agent communication system faces many methodological challenges, especially in the aspect of dynamic system communicative behavior. To alleviate the situation, we develop a suite of methods for analyzing, designing, and verifying communicative behavior within an agent communication system. These methods help us to extract conversation structure from example conversation scenario, help us to check the semantic validity of the communicative act (i.e. Speech-act) occurring the scenario, and help us to verify that different implementation strategies adopted by different agents indeed conform to the protocols they need to comply with. As a whole, our work has followed a software engineering approach to investigating agent systems. Therefore, as a conclusion of our work, we finally apply our languages, models and methods to the analysis and design of a realistic distributed application system. The outcome of the analysis and design is evaluated, and our methods are also compared o many other methods coming from both traditional software engineering field and from the emergent agent-based software engineering field, illustrating their relative merits and drawbacks, and pointing the directions for the future improvement.
语种: 中文
内容类型: 学位论文
URI标识: http://ir.iscas.ac.cn/handle/311060/7280
Appears in Collections:中科院软件所

Files in This Item:
File Name/ File Size Content Type Version Access License
LW004412.pdf(1601KB)----限制开放-- 联系获取全文

Recommended Citation:
王栩. Agent系统通讯理论及组织结构的研究[D]. 中国科学院软件研究所. 中国科学院软件研究所. 2000-01-01.
Service
Recommend this item
Sava as my favorate item
Show this item's statistics
Export Endnote File
Google Scholar
Similar articles in Google Scholar
[王栩]'s Articles
CSDL cross search
Similar articles in CSDL Cross Search
[王栩]‘s Articles
Related Copyright Policies
Null
Social Bookmarking
Add to CiteULike Add to Connotea Add to Del.icio.us Add to Digg Add to Reddit
所有评论 (0)
暂无评论
 
评注功能仅针对注册用户开放,请您登录
您对该条目有什么异议,请填写以下表单,管理员会尽快联系您。
内 容:
Email:  *
单位:
验证码:   刷新
您在IR的使用过程中有什么好的想法或者建议可以反馈给我们。
标 题:
 *
内 容:
Email:  *
验证码:   刷新

Items in IR are protected by copyright, with all rights reserved, unless otherwise indicated.

 

 

Valid XHTML 1.0!
Copyright © 2007-2017  中国科学院软件研究所 - Feedback
Powered by CSpace