Customers scheduling and clustering as vendor managed inventory enablers

Abstract

The vendor managed inventory (VMI) implementation brings advantages to both vendors and customers, improving service levels and transportation efficiency. This paper aims to propose a two stages methodology to enable the VMI implementation without electronic data interchange. First, an algorithm is proposed to determine which customers to attend daily, respecting their needs and restrictions while seeking for economies of scale. Then an iterated local search metaheuristic, adapted to the capacitated p-median problem is proposed to divide the customers into the available trucks. The methodology was applied to a gas distribution case study. Results shown a 20.16% increase in the average quantity of gas delivered per customer, a drop over 90% in delivery delays and a reduction of 31.8% in clusters distances. These results demonstrate that the methodology improves the reliability of the distribution system, ensuring that the VMI implementation brings advantages to all parts involved.