Deploying Neural-Network-Based Models for Dynamic Pricing in Supply Chain Management

With the advent of e-Commerce, enterprises can no longer rely on static business strategies. They have to be able to cope in dynamic and uncertain electronic environments, especially when developing their pricing strategies. In such environments, prices are determined dynamically through a competitive bidding process depending on the market situation, competitor strategies and/or customer preferences. The ability to predict winning bidding prices is crucial. This paper introduces a number of neuralnetwork- based models for performing time-series forecasts of customer offer prices which could result in winning orders in the context of supply chain management (SCM). Different data transformation and normalisation methods are explored in the models, as well as the impact of the number of historical data points included in time-series on the accuracy of the prediction. Experiments in the Trading Agent Competition SCM game show the potential of the proposed algorithms for predicting prices in competitive and dynamic environments. © 2008 IEEE.