In textile industry the production planning troubles the management, due to the variability of textures and fabrics according to the seasons.
Another difficulty is fast fashion, which forces fabric producers to work with push-strategy, selling only product already on stock, with extremely tight delivery times. The main challenge is the identification of colours that need to be in stock, since they depend on fluctuations, with a variability also within the same selling season of few weeks.
The project consists in the realization of an algorithm that, starting from sales, is able to forecast the following weeks orders, with the main goal of automatize the production planning. More in details, the algorithm is making use of hysteresis phenomena typical of buying behaviours, anticipating colours trends and any short-term variation, directly updating the production planning. Artificial Intelligence is going to be the core technology for the algorithm development.
Under a technical point of view, the algorithm will be engineered as a computer container, capable of input and output communication with the most common MES systems.