基于FCM的逆物流供应商评估建模和算法

An Evaluation Model and Algorithm of Reverse Logistics Provider Based on FCM Method

针对闭环供应链环境下逆物流供应商评估的指标及其特点,采用模糊认知图(FCM)方法,建立了逆物流供应商的评估模型。模型结合运用了非线性Hebbian学习算法,该算法通过对认知图的学习训练,避免了评估过程对专家意见的依赖性。最后通过算例说明了本文模型的应用。结果表明,利用模糊认知图建立供应商评估模型是可行的。与AHP方法相比,该方法考虑了标准之间相互影响及反馈的关系,与ANP方法相比,避免了构建超矩阵的复杂性。因此该模型对于企业合理的评估选择逆物流供应商,具有一定的指导意义。

Recurring environmental issues and new regulations are forcing enterprises to improve their reverse logistics to stay profitable.Enterprises could reduce production costs and increase profitability through effective utilization of renewable resources when implementing closed-loop supply chains.By implementing such supply chains,enterprises can develop sustainable resources and economy.A closed-loop supply chain,which covers the whole product life cycle,is required not only by environmental protection and corporate responsibility of external needs,but also by the internal demand for conversion from traditional economy to cycling economy.In order to enhance an enterprise＇s core capability,the enterprise need to outsource reverse logistics as part of a closed-loop supply chain management（SCM） strategy.The main purpose of this paper is to illustrate how to choose a reverse logistics provider properly.Selection of quality reverse logistics providers is extremely important in the effective management of a closed-loop supply chain.This issue must take into consideration several qualitative and quantitative criteria.These criteria address the existing problems associated with the selection of suppliers.The Analytic Hierarchy Process（AHP） and Analytic Network Process（ANP） are two major methods widely used in the section of supply chain partners.Researchers have tried to identify important factors for supplier selection and develop analysis models based on AHP and ANP methods.AHP /ANP methods are excellent for transforming qualitative analysis of supplier selection problems to quantitative analysis.The AHP method assumes that supplier evaluation criteria are mutually independent.However,real life situations warrant against this assumption.The ANP method considers possible dependencies between criteria and the need to carry out a series of pair-wise judgments in order to improve analysis accuracy.The first part of this study discusses a supplier evaluation system for a closed-loop supply chain based on FCM（Fuzzy Cognitive Maps）,and builds an evaluation model for reverse logistics suppliers.Interactive simulation of the FCM method enables decision makers to receive feedback after the data are synthetically analyzed and qualitatively processed.FCM method also combines the Hebbian learning algorithm,which uses the cognitive map to avoid the bias of experts.The second part of this study constructs fuzzy genitive maps to show the causal relationships among criteria and help companies acquire the best suppliers based on the identified relationships.In summary,this paper recommends companies select quality reverse logistics suppliers based on the FCM method.The FCM method differs from the AHP method in the consideration of the independent criteria and feedback.The FCM method can also simplify the complex process of building a supermatrix in the AHP.This paper provides a scientific and comprehensive view of choosing quality reverse logistics providers.