调度Agent任务招投标的粗模糊集建模研究  

A Rough Fuzzy Set Model for Task Bidding of Scheduling Agent

作　　者：包振强[1] 王宁生[1] 蔡宗琰[1] 

机构地区：[1]南京航空航天大学CIMS工程研究中心

出　　处：《中国机械工程》2003年第22期1943-1946,1977,共5页China Mechanical Engineering

基　　金：国家863高技术研究发展计划资助项目(863-511一910-403);江苏省教育厅自然科学基金资助重点项目(01KJB520010)

摘　　要：通过相识集、招标集、投标集和任务集的概念,描述车间调度控制系统中的调度Agent与资源Agent间任务招投标过程模型;基于任务的属性和资源Agent完成任务的成本、质量、负荷和时间等属性,定义论域上的模糊集,将模糊集中的隶属度函数作为粗集的属性,在模糊集上作截集,获得系统的分类知识;收集样本数据,构造并分析决策表,进而获得调度Agent调度决策知识;应用调度知识进行推理,从参与竞标的若干个资源Agent中,选出最适合招标任务的中标者。该方法建立在规范知识库基础上,有利于Agent知识的管理、再学习和更新,据此建立的系统具有通用性和适应性。The process model of bidding for tasks be-tween the scheduling agent and the resource agents incontrol system of shop floor is described in terms of ac-quaintance set, invitee set, bidder set and bid--winner atfirst. Some fuzzy sets on the universe are defined basedon the attributes of tasks and the attributes of the cost,quality, load and time for resource agents to accomplishthe tasks. Considering the membership degrees of thefuzzy sets as attributes of rough sets, the knowledge basecan be obtained by making up a series of cutting sets ofthe fuzzy sets. Through a case study, in this paper deci-sion table is discussed, dependency and significance of at-tributes are analyzed and the scheduling decision rules orscheduling knowledge are deduced. The last part of thepaper aims to instruct a method of fuzzy comprehensivejudgment for the selection of a bid--winner.

关 键 词：粗集 模糊集 AGENT 任务招投标 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程] TP31[自动化与计算机技术—控制科学与工程]