摘要
企业间知识转移是一个复杂的演化过程,当前研究主要集中在理论分析上,通过建模和仿真研究其演化过程的成果尚不多见。从任务需求的角度出发,提出了一种多主体企业间知识转移模型——MASKT(multi-agent system knowledge transfer)模型。该模型引入了强化学习机制及学习对象选择机制,并基于知识收益度对知识转移效果进行了评价。仿真结果表明,强化学习及学习对象选择机制能使企业的知识收益趋于最大化,企业间的知识势差与知识收益的关系呈倒U型结构。
The knowledge transfer is a complex evolutionary process.Currently the research of knowledge transfer focuses on theoretical analysis,the evolution process with modeling and simulation research is still rare.This paper proposed a multi-agent system knowledge transfer model considering the task requirements.This model introduced the reinforcement learning mechanism and learning object selection mechanism,and evaluated the knowledge transfer effect based on knowledge income degree.Simulations show that reinforcement learning and learning object selection mechanism can maximize the knowledge income of enterprise,an inverted U-shaped relationship exists between knowledge distance of enterprises and knowledge income.
出处
《计算机应用研究》
CSCD
北大核心
2011年第10期3756-3759,共4页
Application Research of Computers
基金
山东省科技攻关项目(2009GG10001008)
济南市高校院所自主创新基金资助项目(200906001)
关键词
知识转移
强化学习
多主体
知识收益
仿真
knowledge transfer
reinforcement learning
multi-agent
knowledge income
simulation
