Efficient real-time management of goods distribution to clustered clients

Abstract

This paper focuses on real-time management of a delayed delivery vehicle. The vehicle's initial routing plan is modified to minimise the adverse impact of the delay by selecting the clients to serve and the clients for whom to defer delivery. This single vehicle re-planning problem is modelled based on the formulation of the well-known Orienteering Problem (OP). A novel approach is proposed to provide an efficient, near-optimal solution. This approach decomposes the problem into smaller sub-problems, taking advantage of the geographical distribution of clients into clusters (neighbourhoods and/or suburbs). The available vehicle time is distributed to each cluster by solving a non-linear optimisation problem. Within each cluster, an established OP heuristic is used to determine the clients to be served and the vehicle route. The proposed approach has been applied to large test cases, as well as to problem instances from actual distribution networks. The results indicate that in terms of solution quality the proposed method outperforms the methods used to solve the monolithic problem. Furthermore, the method is, in general, more efficient, especially in cases involving a large number of clients. Finally, computational times are short, indicating the suitability of the method for real-time applications.