A Survey on Metaheuristic Algorithms for Solving Ambulance Routing Problems: Models, Datasets, Tasks and Algorithms

Abstract

Ambulance routing is essential for effective emergency medical service and timely patient transportation within a short period. The ambulance routing problem (ARP) is considered to be one of the challenging fields in the vehicle routing problem with some unique characteristics, such as emergency response, heterogeneous fleets, hospital allocation, patient prioritization, and time-varying demands. These features make it hard to solve optimally using exact solution approaches in reasonable computational time particularly for large-scale instances. Consequently, these characteristics have motivated researchers to apply various metaheuristic approaches. This paper presents a structured survey of ambulance routing problems with different problem models, datasets, research tasks, and algorithms. The majority of the studies have focused on designing models based on the vehicle routing problem framework and reducing response time, travel time, or cost. The most common algorithms used to solve the ambulance routing problem are genetic algorithms, swarm intelligence, tabu search, and hybrid algorithms. However, findings show that neighborhood-based algorithms have demonstrated promising performance for large-scale real-world ambulance routing problems. Additionally, the survey discusses key insights, challenges and directions for future research, such as the need to develop benchmark instances based on real-world data, address the inconsistency in performance, and consider the uncertainty in ambulance routing problems. These insights help researchers and practitioners develop efficient, effective, and reliable optimization frameworks for ambulance routing within the context of emergency medical services. Notably, this survey provides an overview of the connections among problem models, datasets, and algorithms, which can help guide the design of ambulance routing problems and performance evaluation.