Bibliometric analysis and system review of vehicle routing optimization for emergency material distribution

Abstract

Determining the optimal vehicle routing of emergency material distribution (VREMD) is one of the core issues of emergency management, which is strategically important to improve the effectiveness of emergency response and thus reduce the negative impact of large-scale emergency events. To summarize the latest research progress, we collected 511 VREMD-related articles published from 2010 to the present from the Scopus database and conducted a bibliometric analysis using VOSviewer software. Subsequently, we cautiously selected 49 articles from these publications for system review; sorted out the latest research progress in model construction and solution algorithms; and summarized the evolution trend of keywords, research gaps, and future works. The results show that domestic scholars and research organizations held an unqualified advantage regarding the number of published papers. However, these organizations with the most publications performed poorly regarding the number of literature citations. China and the US have contributed the vast majority of the literature, and there are close collaborations between researchers from both countries. The optimization model of VREMD can be divided into single-, multi-, and joint-objective models. The shortest travel time is the most common optimization objective in the single-objective optimization model. Several scholars focus on multiobjective optimization models to consider conflicting objectives simultaneously. In recent literature, scholars have focused on the impact of uncertainty and special events (e.g., COVID-19) on VREMD. Moreover, some scholars focus on joint optimization models to optimize vehicle routes and central locations (or material allocation) simultaneously. Solution algorithms can be divided into two primary categories, i.e., mathematical planning methods and intelligent evolutionary algorithms. The branch and bound algorithm is the most dominant mathematical planning algorithm, while genetic algorithms and their enhancements are the most commonly used intelligent evolutionary algorithms. It is shown that the nondominated sorting genetic algorithm II (NSGA-II) can effectively solve the multiobjective model of VREMD. To further improve the algorithm's performance, researchers have proposed improved hybrid intelligent algorithms that combine the advantages of NSGA-II and certain other algorithms. Scholars have also proposed a series of optimization algorithms for specific scenarios. With the development of new technologies and computation methods, it will be exciting to construct optimization models that consider uncertainty, heterogeneity, and temporality for large-scale real-world issues and develop generalized solution approaches rather than those applicable to specific scenarios.