An application of stochastic programming to inventory management

Abstract

With a continuously growing worldwide economy, commercial industries are facing more competitive markets on a daily basis. Established companies have been investing in warehouses to store the raw materials necessary to build their products in a timely manner to satisfy their ever growing demands. This phenomenon has led to a wide interest in inventory management which has proved beneficial for companies which have made good practice of such warehousing techniques in the past. A local company has shown interest in improving the logistical planning involved in the in-house manufacturing of a new product. The main interest of the company providing the information for the inventory problem that shall be considered, is to minimize the total inventory levels present in the warehouse at any point in time, while simultaneously satisfying their clients’ demands. Since these demands fluctuate with time, this study has delved into statistical demand forecasting and an original stochastic mixed integer linear program has been developed to optimize this complex inventory problem. This has been achieved with the GAMS statistical software using the GUROBI solver where different deterministic reformulations of the program have been solved through a branch-and-cut approach. Solutions to such instances of the proposed program have been reported throughout the study.