演化分析框架下企业学习模型的比较与构建

Comparison and building of firms learning models under evolutionary analytical framework

从演化角度探讨传统主流经济模型的局限,比较分析基于不同角度的学习模型.在此基础上,构建一个基于一定“噪声”条件的随机学习模型,旨在说明建立在学习基础上的企业决策行为的概率演化动态过程.结果表明,演化路径具有复杂系统演化的一般特征,但当系统演化能达到稳定状态时,可以求得一个均衡的决策行为概率分布,这一均衡分布可以作为纳什均衡的随机扩展.

Based on the comparison of the related learning theories from different angles, the paper em- phasizes on the different categories of learning process and compares different learning models, including Bayesian learning models, stochastic learning models, artificial adaptive agent models and models in organization learning theory. On the above basis, the paper models a stochastic learning process of finns, which describes an evolutionary dynamic process of the decision making actions of firms on the probability level. Meanwhile, with the system approaching a stable state, an equilibrium decision making probability density can be derived out, which can be taken as a stochastic generalization of Nash equilibrium.