基于Multi-agent的曲轴随动磨削参数智能决策系统  

Intelligent Decision System of Crankshaft Tracing Grinding Conditions Based on Multi-agent

作　　者：沈南燕[1] 李静[1] 何永义[1] 

机构地区：[1]上海大学CIMS&机器人中心,上海200072

出　　处：《制造技术与机床》2010年第2期162-166,共5页Manufacturing Technology & Machine Tool

摘　　要：为了实现曲轴随动磨削参数的智能选取及优化,采用多Agent技术构建了参数智能决策系统。分别设计了CBR决策Agent、RBR决策Agent、MBR决策Agent、FR决策Agent及相应功能Agent,这四种推理决策Agent的组合能够互相取长补短,避免仅采用某一技术带来的局限。并且建立了曲轴随动磨削参数的范例库、模型库、规则库和方法库以及各库管理系统。针对曲轴随动磨削参数决策特点,制定统一的协调控制策略,并利用黑板解决各Agent间的通信问题,实现了决策Agent之间的有效协同。Multi-agent technology is adopted for building the intelligent decision system to realize the selection and optimization of grinding conditions in crankshaft tracing grinding. The CBR Agent, RBR Agent, MBR Agent and FR Agent, whose cooperation combines the strengths of each individual reasoning approach and overcomes the limitations of any single one, are designed as main modules. The databases of cases, models, rules and methods for crankshaft tracing grinding as well as their relevant management systems are also founded. An uniform control strategy of multi-agent cooperation is established and the blackboard is utilized for the efficient communication among all the agents.

关 键 词：曲轴 随动磨削 多AGENT 磨削参数 决策 

分 类 号：TG58[金属学及工艺—金属切削加工及机床] TP18[自动化与计算机技术—控制理论与控制工程]