MAS不良通信检测方法  

Detection Method of Undesirable Communication in MAS

作　　者：张函[1] 郭锐锋[1] 耿聪[1] 

机构地区：[1]中国科学院沈阳计算技术研究所,沈阳110168

出　　处：《小型微型计算机系统》2015年第5期976-980,共5页Journal of Chinese Computer Systems

基　　金：核高基国家科技重大专项(2012ZX01029001-002)资助

摘　　要：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.Balance of communication among agents is usually the key that affects QoS in MAS. However,the unbalanced states of communication widely exist in the interaction of MAS. Imitating the mechanism operation of human society,the internal behavior characteristics of MAS is considered and the classification method is used for detecting and eliminating undesirable communication structure and unbalanced states of communication, so that the overall QoS can be improved. In this paper, the social characteristic of MAS and the similarity between oppressive behaviors and unbalanced states of communication are firstly discussed. Regarding the oppressive behavior, some metrics have been designed to measure the communication among agents, and some rules have been devised to classify agents into several patterns in the following part. Experiments are conducted on Crisis-MAS in the last. Experiment results show that the causes of the unbalanced communication design can be figured out by means of oppressive behavior and their metrics and classification rules,and the performance atso can be improved by eliminating the oppressive behavior in MAS.

关 键 词：多agent系统(MAS) 压迫行为模式 通信失衡 不良通信检测 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]