Decision surface modelling of apparel retail operations using neural network technology

In this paper the authors develop a metamodel of the relationships between key inputs and performance measures of an apparel retail operation using neural network technology. Two different neural network models, namely a back propagation neural network and a fuzzy control neural network, are investigated. The latter involves the use of a fuzzy controller for the learning rate. Both models are capable of providing high quality predictions, however results obtained using the fuzzy controller suggests a significant potential for speeding up the training process. The resulting neural network is used as the engine in an interactive, graphical management information system.