摘要
供应商是整个供应链的源头,他是企业制造资源的输入端,因而对供应商选择是一个企业经营运作的基础.对供应商的评价选择将在很大程度上决定供应链能否平稳运行和运行的效能.BP神经网络方法可避免传统方法的局限性与专家评价的主观随意性,实现了定性分析与定量分析的有效结合,保证供应商评价结果的客观性,但同时也存在训练时间长、易陷入极小值的缺点.因此,建立了一种神经网络融合技术算法.方法是由主成分分析法、粒子群算法与BP算法相融合而成的一种评价方法.通过收敛性分析和数值模拟,验证该算法具有良好的泛化能力,并且在训练误差、训练时间上也要优于BP神经网络算法.因此,对于解决供应商评价问题是有效的.
The supplier is the source of the entire supply chain, and it is the input of en- terprise manufacturing resources. Thus, supplier selection is the foundation of a business operation.Supplier evaluation is the keystone of the Supply Chain management, evaluation of suppliers will largely determine whether the smooth operation of the supply chain and oper- ating performance. BP neural network method can avoid the limitations of the traditional methods and subjective arbitrariness of expert evaluation .It can make the qualitative analysis and quantitative analysis combination effectively, the evaluation result of the supplier objec- tivity. But BP algorithm also has disadvantages, such as long training time, easily trapped into local minima. In this paper, we design a algorithm with principal component analysis, particle swarm optimization algorithm and BP neural network . By analysis the convergence of the algorithm and numerical simulation experiments we can see that the new algorithm has well application ability, and compared with the BP algorithm, it has small errors and short training time. So the new algorithm for resolve the supplier evaluation problem is valid.
出处
《数学的实践与认识》
北大核心
2015年第24期1-9,共9页
Mathematics in Practice and Theory
关键词
供应商评价
BP算法
主成分分析
粒子群
BP neural network
principal component analysis
particle swarm optimizationalgorithm
