论文部分内容阅读
Multi-agent System(MAS)中agent间通信状态的平衡与否是影响系统整体服务质量的关键因素.在MAS的实际通信过程中,由系统不良通信结构引起的通信失衡状态广泛存在.本研究的目的是利用MAS内部行为特征的仿人类社会群体性,引入压迫行为模式的分类方法来检测和消除MAS中的不良通信结构和通信失衡状态,从而提高系统的整体服务质量.本文首先分析了MAS的拟人类社会特征以及压迫行为与通信失衡现象间的本质相似性,然后提出了基于压迫行为模式的通信状态度量方法和agent角色划分机制,从不同层次,不同角度对系统的整体压迫状态进行评估和分析,最后在Crisis-MAS实验平台上对提出的度量方法进行了实例验证.实验结果说明本文提出的度量方法能够准确的定位引起通信失衡状态的不良通信结构,并通过消除系统中的压迫行为提高了系统的整体服务质量.
Multi-agent System (MAS) agent balance between the state of communication is the key factor affecting the overall quality of service system.In the actual communication process of MAS, communication imbalance caused by poor communication structure of the system exists widely.In this study The purpose of this paper is to use the imitation of human social groups based on MAS’s internal behavioral characteristics and to introduce the classification of oppressive behavior patterns to detect and eliminate the bad communication structure and communication imbalance in MAS so as to improve the overall service quality of MAS.In this paper, The nature of human beings and the similarities between the phenomenon of oppression and communication imbalance, and then put forward the communication state measurement and agent role classification mechanism based on the oppression mode, and evaluate the overall state of oppression from different levels and angles Finally, the proposed method is validated on the Crisis-MAS experimental platform.Experimental results show that the proposed method can accurately locate the bad communication structure that caused the communication imbalance and improve the system’s performance by eliminating the oppression of the system The overall quality of service system.