基于粗集的调度Agent知识库建模研究  

Knowledge Base Model for Scheduling Agent Based on Rough Set

作　　者：包振强[1] 王宁生[1] 李斌[2] 

机构地区：[1]南京航空航天大学机电学院,南京210016 [2]扬州大学信息工程学院

出　　处：《南京航空航天大学学报》2003年第6期671-675,共5页Journal of Nanjing University of Aeronautics & Astronautics

基　　金：国家 8 63 /CIMS基金项目 ( 863 -5 1 1 -91 0 -40 3 );江苏省教育厅自然科学基金 ( 0 1 KJB5 2 0 0 1 0 )资助项目

摘　　要：首先通过相识集、招标集、投标集和中标者的概念 ,描述了车间调度系统中的调度 Agent与资源 Agent间任务分配的招投标关系 ;运用粗集理论 ,基于任务的属性和资源 Agent完全任务的成本、质量、负荷和时间等属性构造决策表 ;在分析了决策表的特征基础上 ,生成调度规则集 ,即调度知识库 ;最后给出车间调度Relations between the scheduling agents and the resource agents are described by the definitions of acquaintance set, invitee set, bidder set and bid gainer at first. Applying the theory of rough set, a decision table is built based on the attributes of the tasks and the attributes of the cost, quality, load and time for resource agents to accomplish the tasks. The characters of the decision table are analyzed and the scheduling decision rules or scheduling knowledge are deduced. According to the definition of the dependence of attributes, the dependence of the decision attribute is analyzed on condition attributes and the significance of each condition attribute is obtained in the condition attribute set. Results show that the knowledge deduced from the data is consistent with knowledge of experts. Finally the bidding process is between the scheduling agents and resource agents and the design scheduling arithmetic of the task bidding is given.

关 键 词：粗集 调度知识库 建模 AGENT 车间调度系统 决策表 

分 类 号：TP182[自动化与计算机技术—控制理论与控制工程] TP311.13[自动化与计算机技术—控制科学与工程]