In real-time distribution operations, it is generally assumed that demand process parameters can be precisely estimated. However, this is rarely the case, particularly for products with a short life cycle. Experiments were done to investigate the performance of various strategies for online distribution operations under disruptions caused by surges in demand patterns at a particular facility. Two sets of strategies are considered: (a) optimization-based strategies, in which an off-line optimization problem is formulated and used to update routing and inventory allocation plans, and (b) fixed-tour strategies, in which a priori sets of routes to provide retailers with recourse actions depending on different degrees of realtime information capabilities for controlling the system are used. These are compared against two benchmark policies. Simulation results for two scenarios (high and low inventory holding cost products) show that strategies that use real-time information to update delivery plans systematically outperform strategies in which routes are not modified after the vehicle leaves the depot. The differences tend to be higher in scenarios of products with low inventory holding costs. For replanning strategies, the benefits of en route plan updates are significant, in particular the possibility of diverting the vehicle. In the case of fixed-tour strategies, the possibility of updating tour intervals provides potentially great savings. Replanning strategies systematically outperform fixed-tour strategies.
ASJC Scopus subject areas
- Civil and Structural Engineering
- Mechanical Engineering