Renovating Intelligent Operations in Supermarket Chains

Abstract

This paper is focused on the development of supermarket models for managing store resources; the proposed approach is based on combining simulation, time series analysis and data fusion for creating forecasts able to support personnel organization. The model operates receiving data provided by the supermarket network; the time series data related to sales, working hour, customers and material flows are processed and then the forecasts are subjected to data fusion and combined with simulation expectation of workloads. The model provides forecasts of the total operative workload for planning activities in supermarkets; in addition the proposed approach allows to estimate future values for target functions such as sales and productivity and to compare historical data in order to adopt predictive resource and policy management. The proposed model is based on a special architecture developed ad hoc by the authors for managing retail networks.This architecture has been implemented using Web technologies providing an easy access for final users to all the predictive models and algorithms. An important added value of this research is related to the identification and definition of new KPI (Key Performance Index) devoted to validate and tune up the model as well as to support management in supermarket. This system was applied to a real case study involving one of the biggest retail company in Italy.