团队是动态不可预测性环境下协作问题求解的有效方式,联合意图是团队联合求解的关键.因此,主体在团队活动中如何采用言语动作形成、维护、解除联合意图,是一个值得研究的重要问题.旨在设计一种基于主体通信语言FIPA(foundation for inte...团队是动态不可预测性环境下协作问题求解的有效方式,联合意图是团队联合求解的关键.因此,主体在团队活动中如何采用言语动作形成、维护、解除联合意图,是一个值得研究的重要问题.旨在设计一种基于主体通信语言FIPA(foundation for intelligent physical Agent) ACL(Agent communication language)的多主体团队交互协议.首先,分析了现有FIPA ACL支持团队联合求解的充分性问题.在概念上明确区分了联合请求与委托请求,指出委托请求言语动作不能充分支持团队协作.并扩展定义了联合请求,讨论了相关定理.然后,基于联合请求动作,提出一种主体团队交互协议,并给出了协议的形式化语义,最后讨论了协议的实际应用.区别于现有的基本动作请求协议、合同网协议以及拍卖协议,团队交互协议描述了另一类主体交互模式,对主体交互模块的设计具有指导意义.展开更多
The behavior decision of agent in multi-agent system is dependent on the task of individual agent ,the cooperation with other agents,and the multi-agent system' constraints. The paper presents three abstract conce...The behavior decision of agent in multi-agent system is dependent on the task of individual agent ,the cooperation with other agents,and the multi-agent system' constraints. The paper presents three abstract concept models of achievement intention,joint intention and maintenance intention representing the factors that affect agent's behaviors respectively,to investigate autonomous agent's behaviors decision in support of agent-oriented software development. The paper discusses and analyzes how they will affect agent's behavior decision and the relationship among them,defines their formal and rigorous semantics,finally specifies and proves a number of important properties.展开更多
The joint intention is an important concept to specify and examine the social behaviors of multi-agent system. A theoretical framework of joint intention is presented. It covers two different joint intentions to repre...The joint intention is an important concept to specify and examine the social behaviors of multi-agent system. A theoretical framework of joint intention is presented. It covers two different joint intentions to represent different social characteristics, including joint achievement intention and joint maintenance intention. The characteristicsof joint intention are investigated, the new semantics of joint intention is given, and several important properties arespecified and proved. The theoretical framework can be used to support the development of multi-agent system.展开更多
An agent-based data mining framework for the high-dimensional environment is built instead of the style of classical structural programming or the object-oriented programming. The framework supports the whole process ...An agent-based data mining framework for the high-dimensional environment is built instead of the style of classical structural programming or the object-oriented programming. The framework supports the whole process of data mining of the high-dimensional environment. Belief-desire-joint intention agents are designed to fit the characteristic of the high-dimensional environment. At the same time, the syntax, semantics and reasoning rules of the agents are given. In the data mining system of the high-dimensional environment, agents need exchange messages. The cooperation behavior mechanism is adopted to complete the communication through the three-level pattern among agents that have their own fixed roles.展开更多
This paper introduces the research of norms and coordination methods for multi-agent system. Finally, we point out some problems and present several research directions, especially the social property of Multi-Agent S...This paper introduces the research of norms and coordination methods for multi-agent system. Finally, we point out some problems and present several research directions, especially the social property of Multi-Agent System.展开更多
文摘团队是动态不可预测性环境下协作问题求解的有效方式,联合意图是团队联合求解的关键.因此,主体在团队活动中如何采用言语动作形成、维护、解除联合意图,是一个值得研究的重要问题.旨在设计一种基于主体通信语言FIPA(foundation for intelligent physical Agent) ACL(Agent communication language)的多主体团队交互协议.首先,分析了现有FIPA ACL支持团队联合求解的充分性问题.在概念上明确区分了联合请求与委托请求,指出委托请求言语动作不能充分支持团队协作.并扩展定义了联合请求,讨论了相关定理.然后,基于联合请求动作,提出一种主体团队交互协议,并给出了协议的形式化语义,最后讨论了协议的实际应用.区别于现有的基本动作请求协议、合同网协议以及拍卖协议,团队交互协议描述了另一类主体交互模式,对主体交互模块的设计具有指导意义.
文摘The behavior decision of agent in multi-agent system is dependent on the task of individual agent ,the cooperation with other agents,and the multi-agent system' constraints. The paper presents three abstract concept models of achievement intention,joint intention and maintenance intention representing the factors that affect agent's behaviors respectively,to investigate autonomous agent's behaviors decision in support of agent-oriented software development. The paper discusses and analyzes how they will affect agent's behavior decision and the relationship among them,defines their formal and rigorous semantics,finally specifies and proves a number of important properties.
文摘The joint intention is an important concept to specify and examine the social behaviors of multi-agent system. A theoretical framework of joint intention is presented. It covers two different joint intentions to represent different social characteristics, including joint achievement intention and joint maintenance intention. The characteristicsof joint intention are investigated, the new semantics of joint intention is given, and several important properties arespecified and proved. The theoretical framework can be used to support the development of multi-agent system.
文摘An agent-based data mining framework for the high-dimensional environment is built instead of the style of classical structural programming or the object-oriented programming. The framework supports the whole process of data mining of the high-dimensional environment. Belief-desire-joint intention agents are designed to fit the characteristic of the high-dimensional environment. At the same time, the syntax, semantics and reasoning rules of the agents are given. In the data mining system of the high-dimensional environment, agents need exchange messages. The cooperation behavior mechanism is adopted to complete the communication through the three-level pattern among agents that have their own fixed roles.
文摘This paper introduces the research of norms and coordination methods for multi-agent system. Finally, we point out some problems and present several research directions, especially the social property of Multi-Agent System.
