基于机器学习的企业定价算法研究

The Enterprise Pricing Algorithm Based on Machine Learning

本文根据经济微观仿真模型ASPEN的定价机理 ,提出了一种自动学习算法。该算法借鉴分类器系统中的规则及消息系统和信任分配系统 ,模拟了企业通过渐进学习得到最大利润下的价格的过程。文中以一个简单经济模型中的两个企业代理人的价值竞争为例 ,将算法的定价与理论计算法的结果进行了比较 。

In this paper, a self-learning algorithm is put forward based on pricing principle of ASPEN, a microsimulation model of the economy. Using rule and message system and apportionment of credit system for reference, the algorithm simulates the course of firms setting price to maximize profit by gradually learning. In this paper, the algorithm is applied to a simple economics model, in which two firm agents carry on price competition, and price obtained using the algorithm is contrasted with the calculated result in theory. The contrast shows that the algorithm is good in simulating the self-learning process of firms setting price.