A Cluster-First Route-Second Constructive Heuristic Method for Emergency Logistics Scheduling in Urban Transport Networks

Abstract

Advanced strategies for emergency logistics scheduling problems in urban transport networks have been a challenging topic for centuries. This study proposed a cluster-first route-second constructive heuristic method based on the continuous approximation (CA) for ‘one-to-many’ vehicle routing to dispatch commidities after an emergency. The objective of the study is to provide a replenish schedule and routing solution from the government/provider’s end in order to minimize the total motion cost, pipeline inventory cost, and holding cost with backorder for the disaster relief operation. The developed method can turn the complicated vehicle routing problem (VRP) into a relatively simple travel salesman problem (TSP) for pre-assigned customer sets. The CA is employed to determine the optimal replenish amount and inventory level for the route serving a given location. The Christofides method is then applied to solve the TSP for the selected cluster. Two clustering methods are investigated in this research: (1) a local-based approach where clustering and routing are determined; and (2) a K-mean clustering method where points are clustered upfront by the CA solution. A case study in Miami-Dade County in Florida to dispatch fuels from the depot to 72 gas stations is presented, demonstrating the proposed approach and comparing two clustering methods. The numerical results illustrate the effectiveness of the algorithms and conclude that the local-based clustering approach may yield a lower total cost with a higher motion cost.