Models and solution methods for the truck routing problem with drone and boat in flood disaster relief distribution response from a sustainability perspective

Abstract

This study proposes three serial studies related to the flood disaster relief distribution response from three pillars of sustainability perspectives, such as economic, social, and environmental pillars. The first study presents an optimization model for distributing logistical items from a warehouse to shelters using the economic pillar aspect. The model utilizes three transportation modes, namely a truck, a drone, and an inflatable boat. This proposed model is called the Traveling Salesman Problem with Drone and Boat (TSP-DB). The truck acts as a mothership vehicle, carrying a drone and a boat. Shelters in the dry area can be served by a truck or a drone, while those in the flooded area can be accessed by a boat or a drone. The drone and boat are deployed from the truck to deliver items to shelters. Due to the limited capacity and the high relative demand at a shelter, the drone can only make one visit at a time before returning to the truck, while the boat can perform multiple visits in a single trip. The objective is to obtain the efficient route by minimizing the completion time. The proposed problem is first modeled using mixed-integer linear programming. As the problem is hard to solve exactly, especially for relatively large-sized datasets, an effective matheuristic that combines an exact method and heuristic, called a Matheuristic-based Dynamic Record-to-Record Travel Algorithm (MDRRT), is then proposed. The performance of the MDRRT is assessed using generated and benchmark datasets. The results reveal that the proposed method is robust and competitive when compared to existing state-of-the-art methods on related traveling salesman problems with drone models. The MDRRT is also applied to a realistic case study, where interesting and valuable managerial insights are discussed and analyzed.
If social conflict among victims is vulnerable to occur, then humanitarian logistics should be focused on the social aspect in addition to the economic, which is usually called an equitable aspect. Therefore, the second study aims to develop the Bi-objective TSP-DB (BTSP-DB) for disaster relief distribution by utilizing truck-drone-boat vehicles with an equity-based perspective. Moreover, load-dependent vehicle speed is also applied in this model to imitate the realistic movement of vehicles in distributing the logistical items during disaster response. Novel non-linear and linear optimization models for the truck drone-boat routing problem are developed, which are then solved using an ε-constraint method (EC). This study also proposed the matheuristic-based speedup mechanism to reduce the computational time of EC, called an ε-Constraint with Dynamic Hill Climbing based method (EC-DHC). The computational results on hypothetical data of the case study and generated datasets show that the proposed method produces a promising result to tackle the bi-objective problem, followed by providing a valuable managerial implication related to vehicle-tandem utilization, policies, and equity scenarios that can be executed to increase the effectiveness of disaster relief distribution.
In the annual flood disaster management context, a new potential flooded region may be detected. Stakeholders should develop a contingency plan to handle the flood disaster response in the new area that can be sustained for a long period. Therefore, the third study presents a sustainable flood disaster response framework with two interrelated and complex phases of optimization. In the first phase, the Multi-objective Shelter and Hub Location Problem (MSHLP) is proposed, which aims to find the best locations for lo cating shelters and a hub by considering three pillars of sustainability. A hybridization of a Genetic algorithm (GA) and an exact method, referred to as a Matheuristic-based GA (MGA), is designed to generate non-dominated solutions. The quality of each non dominated solution is then evaluated using the second-phase optimization model by ad dressing the routing problem for delivering humanitarian aid from a warehouse to selected shelters, called the Mothership and Drone-Boat Routing Problem (MDBRP). This routing problem also considers quick response teams that need to be allocated to help victims in the shelters. This study put forward the combination of a Variable Neighborhood Search (VNS) and an exact method to solve the routing problem. Moreover, this study also en hances the method with a dynamic threshold-based procedure to accept inferior solutions to avoid getting trapped in the local optimal solution. This proposed method is called a Matheuristic-based VNS with Threshold (MVNST). The performances of MGA and MVNST are tested using generated and case study datasets. The results show that the proposed models and methods have promising results, where applicable managerial impli cations are also obtained. To sum up, the three serial studies provide valuable managerial implications for handling a sustainable flood disaster relief distribution response.

Date of Award	15 Oct 2025
Original language	English
Awarding Institution	University of Nottingham
Supervisor	Chandra Irawan (Supervisor) & Zhao Cai (Supervisor)