Research On Enterprise Supplier Evaluation Based On The Adaptive Neuro-fuzzy Inference System

With the increasing competition of the global market and the further introduction of modern management theory,the enterprise's competitive advantage is increasingly dependent on the support of the external organizations,instead of just their own capabilities.The supply chain management has become the core element of the enterprise competitiveness.As the source of the supply chain,the importance of supplier evaluation is obvious.So how to realize the scientifically accurate and simply practical evaluation of the enterprise is a problem worthy of study.In this paper,we design an adaptive neural network system based on the reviewing of pertinent literature and books on the supplier evaluation,fuzzy theory and artificial neural network.Firstly,under the premise of a complete supplier evaluation index system,dimensionality reduction is conducted with the factor analysis method,which also can ensure the original data information as little as possible.Secondly,combining the fuzzy logic and neural network,the adaptive neuro fuzzy system is constructed.The five layer structure system can be used to realize supervised learning by using the hybrid learning algorithm,which integrate the least square algorithm and the gradient descent method algorithm.And finally with the above steps,it realizes the intended functionality of the supplier evaluation and selection for enterprise.In the end,the paper uses the supplier data of the multinational company R to verify the evaluation system.As the empirical analysis result show,the system performs its good accuracy when it comes to new supplier evaluation.Moreover,with both the clear structure and convenient calculation,the system possesses practical value.