In a busy supermarket the number of checkout lanes is constant. Not all customers buy in large volumes. Some buy in small quantity but are forced to wait in long queues at the checkout counter. Our objective is to demarcate a few checkout lanes as "fast checkout lanes" which would exclusively serve these customers (who shop in low quantity), thereby lowering their waiting time. The challenge however is to predict an optimum number of fast checkout lanes, such that on one hand they are able to process these customers fast and on the other they do not remain empty.
The fast checkout lanes, in this case will be altered dynamically everyday based on the predicted demand i.e. the percentage of fast checkout lanes will be directly proportional to the percentage of "small baskets". Hence a model is created which predicts the number of customers with small basket size on a particular day of the week. The model should take into consideration the weekly demand cycle as well as the seasonal variation.

Benefit – The benefit of optimizing the checkout lanes is that it improves the service levels and improves customer satisfaction by reducing the time spent waiting. The optimization also seeks to balance the load on fast and regular checkout lanes.