Online simulation for a real-time route dispatching problem. http://dx.doi.org/10.1057/jors.2011.151

Abstract

In this article we present an online simulation application for a decision problem that operates in realtime, where products have to be dispatched from two depots to clients that are geographically distributedthroughout the city. The system’s behaviour is highly stochastic, due to the random behaviour of theclient’s demand (in time and space), and the random times of order preparation, travelling times ofdispatchers (these are motorcycle drivers) and absence rate of drivers each day. A decision scheme isproposed that combines elements of vehicle routing with time windows, real-time dispatching of driversand online simulation, through which information on future events is considered in the decision-makingprocess. Two major conclusions are obtained when this scheme is applied to real data. First, we showthat the proposed algorithm for order consolidation and route dispatching can be very advantageousfrom the point of view of logistics costs and quality of service. Second, we show that online simulationand, specifically, the Simulation-based Real-time Decision Making methodology (SRDM) can furtherimprove the quality of the results. New ideas for further work are also proposed.