摘要
提出了具有较高稳定性的两级案例推理方法,研究如何采用框架结构表达钢铁生产动态调度的案例.引入决策树的思想实现案例聚类,运用最邻近法实现案例检索.借鉴面向对象的思想调整案例,采用模糊集理论实现案例评价,并探讨了案例维护的基本方式.开发了原型系统,运行结果在速度、有效性和稳定性方面均符合现场基本要求.
The two-stage case-based reasoning (CBR) is proposed to realize stability of the dynamic real-time scheduling system of iron and steel works. The frame structure presentation is used to describe complex scheduling problems in iron and steel production. By introducing decision tree idea to cluster cases, the nearest neighbor (NN) algorithm is used to retrieve case. The techniques such as adopting O-O idea to adapt case, evaluating cases by the fuzzy theory, and the primary method of case maintaining are discussed. The computation experiment shows that the developed prototype system can satisfy the requirement of practical production on velocity, validity and stability.
出处
《控制与决策》
EI
CSCD
北大核心
2004年第10期1109-1112,1116,共5页
Control and Decision
基金
国家自然科学基金项目(70171056
60084003)
国家863/CIMS项目(2002AA412010
2002AA414610).
关键词
案例推理
动态调度
检索
决策树
模糊集理论
Decision making
Fuzzy sets
Information retrieval systems
Real time systems
Scheduling
