面向LDSS的回程配货网点SOFM选址建模方法

针对我国第三方物流企业回程配送作业效率普遍不高的问题，从物流企业回程物流设施网络构建角度出发，运用人工神经网络自组织理论，分析了回程配货的特点，构建了回程配送网点SOFM选址模型.结合实例进行方法应用，运用Witness仿真软件对所建立的模型在有无回程配货网点的两种情景下进行仿真分析，验证了科学合理的回程配货网点可以支持企业提高运送效率并减少运作成本，具有一定的可行性，为构建LDSS提供了理论基础准备.

SOFM location modeling method of LDSS oriented returning goods distribution network

In order to solve the low efficiency problem in the returning goods distribution of third party logistics enterprises in our country, the characteristic of the returning goods distribution was analyzed and the self organizing feature map （SOFM） location model for the returning goods distribution network was established based on artifical neural network self organizing theory and from the view of the returning logistics facility network building. The method was applied in the practical examples. The established model was simulated and analyzed using Witness simulation software in two cases with and without the returning goods distribution network. It is verified that the scientific and rational returning goods distribution network is favorable for improving the distribution efficiency and decreasing the operation cost. The proposed modeling method is of certain feasibility and provides a theoretical basis for establishing the logistics decision support system （LDSS）. 