Abstract
In the process of industrial gas distribution, the uncertainty of future demand presents an important challenge for the production of efficient delivery plans, where the overall cost is minimized and product shortages must be avoided. The objective of this paper is to present a new method for producing future delivery plans that eliminate stock-out risk and minimize transportation and inventory holding costs, in the face of time window constraints. The proposed method follows a Economic Model Predictive Control (EMPC) approach, which makes use of a new mathematical programming model of the distribution process to predict future inventory levels. The method is illustrated using an example with a simple forecasting approach. It is shown that the stock-outs risk can be eliminated, but still balanced decisions between cost saving and risk aversion remain at the discretion of the planner. It is also shown that coordinated inventory routing and control through the EMPC approach brings economic benefits and hedges effectively against the market volatility, while opting for minimum overall costs may incur excessive shortage risk.
