基于多agent的多级供应链需求预测研究

The Research on Demand Forecast of Multi-echelon Supply Chain Based on Multiple Agents

用多agent系统模拟一个多级供应链,通过各个agent来运行供应链中的各个实体,构建了基于多agent系统的多级供应链模型;在此基础上利用遗传算法预测需求,计算供应链中各个实体的需求量;并和其它需求预测方法进行对比;对比结果表明,该方法优于移动平均值、指数平滑等需求预测方法。

A multi-echelon supply chain was simulated by a multi-agent system, where agents operate these entities with different inventory system in supply chain, A model about multi-echelon supply chain was set up based on a multiagent system. In the model, the method of demand forecasting was used by genetic algorithm. The ordering quantity of every entity was offered by the method. Comparing genetic algorithm with moving-average forecasting and exponentialsmoothing forecasting methods, the result shows that the genetic algorithm forecasting method is superior to others.