Distribution Hub Optimization: Application of Conditional P-Median Using Road Network Distances

TL;DR

This paper presents a GIS-based algorithm for optimizing last-mile delivery hubs using road network distances, demonstrating a 16% reduction in delivery distances in Pakistan's logistics context.

Contribution

It introduces a novel application of the conditional p-median problem with road network distances for hub placement in last-mile logistics.

Findings

Achieved a 16% reduction in average delivery distance.

Developed an algorithm utilizing road-network-based distances.

Validated the approach with real and simulated data.

Abstract

This paper explores a GIS-based application of the conditional p-median problem (where p = 1) in last-mile delivery logistics. The rapid growth of e-commerce in Pakistan has primarily benefited logistics companies, which face the challenge of resolving inefficiencies in the existing infrastructure and scaling effectively to meet increasing demand. Addressing these challenges would not only reduce operational costs but also lower carbon footprints. We present an algorithm that utilizes road-network-based distances to determine the optimal location for a new hub facility, a problem known in operations research as the conditional p-median problem. The algorithm optimizes the placement of a new facility, given q existing facilities. The past delivery data for this research was provided by Muller and Phipps Logistics Pakistan. Our method involves constructing a distance matrix between candidate hub locations and past delivery points, followed by a grid search to identify the optimal hub location. To simulate the absence of past delivery data, we repeated the process using the population distribution of Lahore. Our results demonstrate a 16% reduction in average delivery distance with the addition of a new hub.