Multiple Forecasting Algorithms for Demand Forecasting in the Fashion Industry

Due to the global economic crises, demand forecasting is becoming a critical issue to assure high level customers' service and strategies effectiveness. This is even more important for the fashion industry were forecasts are used both for yearly demand estimations and for real-time forecasts. Yearly estimations are usually important because it is difficult in terms of production rate to face demand deviation in short time and, to this end, they allow increasing the responsiveness of the system. Also real time estimations are important for implementing better transshipment policies between stores belonging to the same supply chain echelon. In this paper the problem of demand forecasting in the fashion industry is approached from a different point of view. In particular, a forecasting model based on the use of multiple autoregressive algorithms and disaggregation policies is proposed. A case study showcases the validity of the proposed algorithms by using a Modeling & Simulation based approach.