Running a fast-food restaurant requires a high degree of planning skill, as speed and cost efficiency play a major role in this situation. On the one hand, a fast-food restaurant is forced to keep the waiting times after ordering to a minimum whilst, on the other hand, product quality must be seen to be believed in order to achieve a lasting customer bond. In practice this is solved by means of a compromise, whereby products are prepared and kept in reserve. Nevertheless, in order to guarantee the quality of the products, the products are only kept in the chutes known by everyone for a certain time. If they are not sold within the maximum storage time, then the products are disposed of.
This presents a challenge from a planning point of view, as the person doing the planning has to estimate, how many products have to be prepared and kept in reserve in order to maintain the maximum level of efficiency with the minimum amount of waiting time. For these reasons, the GECO►C team statistically analysed the sales data of the products. The aim was to develop a system that will produce a recommendation with regard to the production of supplies using past sales figures and taking into account other factors having a bearing, such as the weather, the day of the week or special promotions.
In order to solve the task, artificial neural networks (ANNs) were used in combination with statistical processes. The neural networks ensured that the information contained within the sales figures with regard to purchasing behaviour in the past was extracted in order to produce a forecast for the production of supplies from it. The results achieved showed a clear match between the sales figures forecast and those achieved in reality.
