Emergency logistics network design based on space–time resource configuration

https://doi.org/10.1016/j.knosys.2021.107041Get rights and content

Highlights

  • Study a collaborative emergency logistics network based on resource sharing.

  • Propose a bi-objective optimization model for emergency logistics delivery.

  • Devise a hybrid algorithm with k-means clustering and CW_NSGA-II to solve the model.

  • Conduct a case study to show the applicability of proposed model and algorithms.

Abstract

The occurrence of natural disasters or accidents causes the obstruction or interruption of road traffic connectivity and affects the transportation of essential materials, especially for cross-regional delivery under emergency situations. Affected by COVID-19, government administrators establish cross-regional quarantine roadblocks to reduce the risk of virus transmission caused by cross-regional transportation. In this study, we propose an emergency logistics network design problem with resource sharing under collaborative alliances. We construct a state–space–time network-based bi-objective mixed integer programming model to optimize the vehicle routes in order to meet customer demands for essential materials with the lowest cost and highest emergency response speed under limited transportation resources. A two-stage hybrid heuristic algorithm is then proposed to find good-quality solutions for the problem. Clustering results are obtained using a 3D k-means clustering algorithm with the consideration of time and space indices. The optimization of the initial population generated by the improved Clarke and Wright saving method and improved nondominated sorting genetic algorithm-II with elite retention strategy provides stable and excellent performance for the searching of Pareto frontier. The cost difference of the entire emergency logistics network before and after collaboration, i.e., the profit, is fairly allocated to the participants (i.e., logistics service providers) through the Shapley value method. A real-world case in Chongqing City, China is used to validate the effectiveness of the proposed model and algorithm. This study contributes to smart transportation and logistics system in emergency planning and has particular implications for the optimal response of existing logistics system to the current COVID-19 pandemic.

Introduction

The construction of an emergency logistics network is a necessary and crucial task for government officials and managers to protect the residents’ normal life in the occurrence of some natural disasters or accidents ([1]). At the end of 2019, the outbreak of COVID-19 has caused many countries in the world to enter a state of emergency. Several countries have imposed some strict traffic control policies, including the policy of closing cities, to deal with the high infectivity and mortality of the new coronavirus. On the one hand, measures, such as road traffic restrictions, including cross-regional quarantine inspection of transportation vehicles and prohibition of residents from purchasing materials across regions, reduce the risk of virus transmission, thereby ensuring the health of residents. On the other hand, ensuring the accessibility and timeliness of transportation of the living supplies is an important issue that must be addressed by logistics service providers and governors when transportation resources are limited and transportation is blocked.

Compared with the logistics networks under a non-emergency mode, more factors, such as emergency response time, limited transportation resources, and high transportation costs, need to be considered in emergencies ([2], [3]). Sudden disasters, such as earthquakes, epidemic outbreaks, and other emergencies, usually have a series of adverse effects on logistics delivery activities, such as the lengthened travel time, the shortage of transportation resources, and the obstruction of cross-regional delivery operations. However, the timely and safe delivery of living materials to urban residents is vital to ensuring social stability. As a government manager, increasing the response speed to material delivery under emergency conditions with limited resources is necessary to ensure the normal delivery for residents. From an enterprise’s perspective, logistics service providers need to optimize delivery networks under various policy constraints to ensure the normal operation of their enterprises with a relatively low cost. Therefore, an optimization strategy that considers multiple decision goals is needed to achieve a holistic design of the delivery network in an emergency.

The design of emergency logistics network for living materials plays a vital role in maintaining normal social order and ensuring residents’ daily lives when emergencies or unexpected disasters such as earthquakes and epidemics occur ([4], [5])). The occurrence of an emergency involves the research and exploration of the following issues: transportation supplies of living materials, formulation of evacuation plans, placement and transfer of casualties, and the location of emergency logistics facilities ([6], [7], [8], [9]). In the past, research on emergency logistics and discussions and analyses for actual disasters are very limited due to the accidental and the low-frequency characteristics of disasters ([10], [11]). Recently, with the increasing awareness of humanitarian and human fate community, more and more academic experts and decision-makers of enterprises have focused on the research on emergency logistics ([12], [13]). For example, in 2020, with the impact of COVID-19, the time and cost of cross-regional transportation are inevitably increased to avoid the risk of virus transmission caused by cross-regional contact due to quarantine, epidemic prevention, and other policies ([14]). Therefore, designing a safe and reliable emergency logistics network in face of the traffic interruption, such as the increase in travel time or costs for some sections of cross-regional transportation is of vital importance to social stability ([15], [16], [17]).

Emergency logistics network design has focused on the optimization of relief routing in previous literature, especially on the quick and safe transfer of people who have experienced disasters to a predetermined refuge after a disaster ([18], [19]). The study of emergency logistics network design aims to find an effective solution for distributing life supplies to customers in need. Some researchers considered an emergency logistics network design in a deterministic environment with objectives of minimizing total delivery time ([20], [21], [22]). In response to emergency situations that may cause road damage such as natural disasters, Yan and Shih [23] added emergency road maintenance considerations to solve a combined emergency logistics network optimization problem. In addition, delivery efficiency and operational cost should be taken into consideration when designing the emergency logistics network to ensure fairness ([24]). However, for the possible road interruptions after disasters, such as the blocking of roads by debris in the emergency logistics network, many scholars focused on removing road barriers with minimum cost or minimum cleaning time to ensure the connectivity and reachability of the emergency delivery network Celik et al., Sahin et al., Berktas et al., Akbari and Salman [25], [26], [27], [28]. The research on how to satisfy the daily customer demands on the basis of existing accessible roads through multi-facility collaboration needs further exploration. Therefore, we consider the design of the emergency logistics network based on the feasible transportation network, and establish a collaborative delivery mode to meet the delivery demands of customers while ensuring the economic benefits of the logistics enterprises.

However, all the aforementioned studies focused on the logistics network design problem for one logistics company only. They did not consider the possible collaboration and resource sharing among multiple logistic service providers, although it is important especially in an emergency. Through the formation of collaborative alliances, we can achieve a win-win solution for both customers and enterprises themselves. In the previous literature, the importance of coordination or allocation of limited resources, especially for multi-facility cross-regional collaborative delivery, has been largely ignored ([29], [30], [31], [32]). Many studies on resource allocation focused on route optimization only within a logistics company to achieve transportation accessibility ([33], [34]). Delays or undeliverable phenomena may occur in some cross-regional long-distance deliveries due to road traffic obstruction and non-collaborative logistics operation modes.

Moreover, most studies considered only one objective function, i.e., minimizing the total delivery cost. However, in an emergency, other aspects are also important. For example, the speed of emergency response, especially the timeliness of emergency delivery of blood, medical resources, daily necessities, etc., is crucial to the lives of residents. Zhou et al. [35] proposed a multi-period emergency resource scheduling problem that aims to minimize the risk of unmet needs and the choice of damaged roads. Shin et al. [1] proposed a comprehensive optimized scheduling strategy for maintenance personnel and rescue vehicles after a disaster, with the goal of minimizing the service completion time of the total demands. Although the timeliness of transportation or delivery after disasters or emergencies is extremely important, especially disasters that may cause casualties, such as earthquakes and hurricanes ([35], [1]), multiple objectives, such as emergency response time, delivery costs, and the efficiency of the used transportation resources (e.g., transportation trucks or vehicles) under emergency conditions are required to be considered simultaneously ([3]). Therefore, we consider addressing a bi-objective optimization problem that takes into account the logistics cost of the enterprise, the satisfaction of customer demands, and the limitation of social available resources to alleviate or improve the possible conflicts in the emergency situation.

Therefore, reducing cross-regional conditions through collaborative logistics operations while satisfying the demand of existing customers is another important issue for emergency logistics network optimization. Most of the previous logistics delivery issues concerning collaboration only considered the vehicle routing optimization through the collaborative operations between multiple facilities in the entire static or dynamic situation rather than the particularity of emergency situations, such as the increase in travel time or costs for some sections of cross-regional transportation ([15], [16], [17]).

To close the above research gaps, we design an emergency logistics network based on a collaborative mode with resource sharing from multiple perspectives, including the cost, the emergency response time, and the effective utilization of transportation resources. The change in cross-regional travel time due to unexpected conditions is considered when calculating the objective functions of the entire delivery network cost and response time. Designing a relatively closed regional collaborative delivery network reduces the transportation time, including quarantine time when transportation trucks perform delivery tasks across regions. The logistics cost of independent operation under noncollaborative mode is also reduced, thereby ensuring the timeliness of delivery services and accessibility of cargo transportation under an emergency situation. The cost difference due to collaboration, that is, the additional profit can be fairly allocated by comparing the total logistics cost under the optimized delivery strategy and the original delivery mode to promote the formation of the collaborative relationship and maintain its stability and sustainability.

Compared with previous research on the optimization ofemergency logistics networks, this study has the following innovations: (i) we propose a collaborative emergency logistics strategy among multiple logistics facilities based on state–space–time network using existing road resources to fulfill the delivery demands rather than the independent operations among different logistics facilities or enterprises; (ii) we consider multiple objectives, including the total logistics delivery costs and the total time, to get the lowest cost and highest emergency response speed while using limited transportation resources, thus achieving the Pareto optimality under multiple objectives; (iii) we adopt a two-stage heuristic algorithm to validate the advantages of the collaborative delivery strategy in an emergency in terms of total delivery cost, total time, and used transportation resources in a real case in Chongqing City, China.

The remainder of the paper is organized as follows. Section 2 presents the problem statement. Section 3 introduces the mathematical model formulation for the investigated problem, including the assumptions and notations. Section 4 proposes a two-stage heuristics algorithm by combining the Clarke and Wright (CW) saving method and improved nondominated sorting genetic algorithm-II (NSGA-II). Section 5 conducts the numerical experiments in Chongqing City, China. Section 6 summarizes the conclusions.

Section snippets

Problem statement

Taking the conventional logistics delivery network in a non-emergency mode as a reference, we explain the superiority of a collaborative emergency logistics network based on resource sharing when an unexpected event occurs. The state–space–time network is used to describe and analyze the emergency response speed and logistics cost between the non-emergency and emergency modes ([36]). Compared with the traditional delivery network optimization, delivery activities are regarded as single

Optimization model

The shortage of transportation resources and the extended travel time for quarantine in emergency conditions may cause the increase in waiting time and logistics operating costs of logistics companies during the delivery process. Therefore, a bi-objective optimization model with the consideration of emergency response speed and transportation resource sharing is conducted to study the design of emergency logistics network, thus to achieve the optimization of total logistics operating costs,

Solution methods

To solve the proposed mathematical model with two objectives and two echelons based on the emergency logistics network, we design a two-stage solution algorithm, including a 3D k-means clustering algorithm based on the enclosed area of the customers’ geographic locations and service time windows and an improved NSGA-II algorithm. The flowchart of the solution framework is shown as Fig. 5.

In Fig. 5, we use the 3D k-means clustering algorithm to divide an enclosed collaborative delivery area. The

Algorithm comparison

The effectiveness and stability of the improved NSGA-II with elite strategy are verified in searching the Pareto frontier of the bi-objective optimization problem. We obtain an improved Solomon dataset by replacing the coordinates in the original Solomon dataset ([48]) with geographic coordinates based on the actual network. We compare and analyze the performance with the traditional NSGA-II ([49]) and multi-objective harmony search algorithm (MOHSA) ([50]) in solving multi-objective problems

Conclusions

This study focuses on the optimal design of emergency logistics network in the face of a natural disaster or some unexpected events considering the multi-facility collaboration and multiple objectives. To address the problem, we first establish a state–space–time network-based mixed-integer programming model to characterize the basic operating mode and optimal design of a two-echelon emergency logistics network. Due to government macro-control and financial subsidies, cross-regional

CRediT authorship contribution statement

Yong Wang: Conceptualization, Methodology, Software, Validation, Formal analysis, Investigation, Data curation, Resources, Visualization, Writing - original draft, Writing - review & editing, Supervision, Project administration, Funding acquisition. Shouguo Peng: Software, Validation, Formal analysis, Investigation, Data curation, Resources, Visualization, Writing - original draft, Writing - review & editing. Min Xu: Conceptualization, Methodology, Formal analysis, Investigation, Data curation,

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgments

The authors would like to express our sincere appreciation for the valuable comments made by three anonymous reviewers, which helped us to improve the quality of this paper. This research is supported by National Natural Science Foundation of China (Project No. 71871035), Humanities and Social Science Youth Fundation of Ministry of Education of China (No. 18YJC630189), Key Science and Technology Research Project of Chongqing Municipal Education Commission (No. KJZD-K202000702), Key Project of

References (53)

  • WangY. et al.

    Green logistics location-routing problem with eco-packages

    Trans. Res. Part E

    (2020)
  • YanS. et al.

    Optimal scheduling of emergency roadway repair and subsequent relief distribution

    Comput. Oper. Res.

    (2009)
  • Camacho-VallejoJ.F. et al.

    A bi-level optimization model for aid distribution after the occurrence of a disaster

    J. Cleaner Prod.

    (2015)
  • SahinH. et al.

    Debris removal during disaster response: a case for turkey

    SOCIO-Econ. Plann. Sci.

    (2016)
  • BerktasN. et al.

    Solution methodologies for debris removal in disaster response

    EURO J. Comput. Optim.

    (2016)
  • AkbariV. et al.

    Multi-vehicle synchronized arc routing problem to restore post-disaster network connectivity

    European J. Oper. Res.

    (2017)
  • WangY. et al.

    Profit distribution in collaborative multiple centers vehicle routing problem

    J. Cleaner Prod.

    (2017)
  • WangY. et al.

    Collaboration and transportation resource sharing in multiple centers vehicle routing optimization with delivery and pickup

    Knowl.-Based Syst.

    (2018)
  • HuangM. et al.

    A continuous approximation approach for assessment routing in disaster relief

    Transp. Res. B

    (2013)
  • BalcikB.

    Site selection and vehicle routing for post-disaster rapid needs assessment

    Trans. Res. Part E

    (2017)
  • ZhouX.Y. et al.

    D-DEMATEL: a new method to identify critical success factors in emergency management

    Saf. Sci.

    (2017)
  • MahmoudiM. et al.

    Finding optimal solutions for vehicle routing problem with pickup and delivery services with time windows: a dynamic programming approach based on state–space–time network representations

    Transp. Res. B

    (2016)
  • GovindanK. et al.

    Two-echelon multiple vehicle location-routing problem with time windows for optimization of sustainable supply chain network of perishable food

    Int. J. Prod. Econ.

    (2014)
  • WangY. et al.

    Collaborative two-echelon multicenter vehicle routing optimization based on state-space–time network representation

    J. Cleaner Prod.

    (2020)
  • DefrynC. et al.

    A fast two-level variable neighborhood search for the clustered vehicle routing problem

    Comput. Oper. Res.

    (2017)
  • QiuM. et al.

    A tabu search algorithm for the vehicle routing problem with discrete split deliveries and pickups

    Comput. Oper. Res.

    (2018)
  • Cited by (0)

    View full text