A recursive neural network model for supplier selection based on genetic algorithm in supply-chain design

In supply chain design, the effective supplier selection and evaluation is regarded as the key process. This paper quotes seven factors as the selection criterions and shows a dynamic artificial neural network(ANN) model with recursive characters. The topology structure of this ANN model is optimized by the genetic algorithm on the decision of network layers and units per layer. This revised model is developed as a DSS software and trained by 15 years' relative data of a native bearing manufacturing company with the topology network structure of 2 hidden layers(3 units for first layer and 2 units for the second).