Package Delivery Route Optimization with Tight Travel Time

Abstract

In the package delivery and courier services industry, delivery time is one crucial factor that influences customer satisfaction. While most delivery package service companies implement hub-and-spoke network design to achieve economies of scale, same-day delivery services between every origin-destination pair can be ensured by designing a delivery network with tight travel time constraints. This study focuses on designing method that could answer the main decisions in hub-and-spoke network design-which are: the optimal number and locations of hubs, along with the allocations of other nodes to hubs-and to transform such hub-and-spoke network design into a routed one which fulfills both tight travel time and minimization of vehicles needed. Uncapacitated Single Allocation p-Hub Median Problem and K-Means clustering method were used to design the initial hub-and-spoke network. The directly linked network then transformed into a routed network by implementing the Local Search algorithm and an Integer Programming model. The optimal network design was chosen by considering the minimum number of vehicles needed. The method was then tested using a case study of a package delivery start-up company operating in Jakarta. Travel time data were collected between every delivery origin-destination pair. Results from both methods shown that 3 is the optimum number of hubs.