Exploring the Applications of Agent-Based Modeling in Transportation

Agent-based modeling (ABM) has emerged as a distinct and innovative paradigm offering an alternative to conventional modeling techniques which often rely on equation-based representations to depict specific events or phenomena.

The transformative nature of agent-based modelling lies in its focus on characterizing the subject of inquiry through the lens of individual agents and their intricate interactions [1]. Over the course of its evolution, agent-based modelling has witnessed a multitude of applications spanning various domains [2], transportation, nevertheless, being one of them.

Transportation stands as a fundamental pillar of our contemporary society, serving as the essential conduit for facilitating the seamless transfer of commodities, individuals, and animals across different locales, encompassing modes such as air, sea, road, rail, cable, pipeline, and even space [3,4]. The maturation of agent-based modelling techniques has facilitated a range of advancements within the realm of transportation, as evidenced by their practical implementations across diverse domains within the transportation sector.

The advantages of using ABMs in the area of transportation are particularly useful for modeling various aspects due to the ABM ability to capture various individual behaviors in complex environments characterized by the occurrence of spatial interactions. Using ABM, one can set rules of behavior to the agents that operate at a micro level, but which, through simulation, can exhibit complex and emergent behavior at a macro level. For example, one can observe the formation of the traffic congestion patterns by simply studying the ABM, which cannot be achieved through the use of the traditional models.

Furthermore, due to the inherent spatial nature of the transportation systems in general, ABMs provide an advantage over the classical modelling approaches as they consider the spatial context, allowing the interaction of agents with one another and with the environment. As a result of these interactions, emergent behavior can be easily observed by simply watching a simulation run.

The underlying mechanisms driving transportation dynamics can be examined in detail through an ABM approach as the agents can change their behaviors over time through various adaptation mechanisms. These changes can be easily captured in an ABM. Various disruptions, such as delays in the schedule of various means of transport or the unavailability of one or more exit doors in a transport evacuation simulation, can be also captured in an ABM.

Nevertheless, the “what if” scenario development option offered by ABM can be useful for testing various policies and scenarios before using them in practice. For example, in the airplane boarding problem, the test for the various boarding methods can be performed through the use of ABM for various quantities of luggage brought by the passengers inside the aircraft.

The applications stemming from the utilization of ABMs within the realm of transportation have ignited a diverse array of domains, encompassing yet not confined to the ones discussed below.

Within the domain of air transportation, the application of ABMs has predominantly centered around the following focal points of research: study of the airport terminal operations [5], estimating occupant density in airports [6], simulating the passenger flow in airport terminals [7], improving the airplane boarding process [8], improving passenger service time [9], assessing airport terminal resilience [10], analyzing the security and efficiency in airport terminals [11]. Among the delineated sub-areas within air transportation, the domain of airplane boarding has witnessed a pronounced surge in applications utilizing agent-based modeling [12]. Issues arise such as evaluating the efficiency and sustainability of airplane boarding strategies, simulating airplane boarding through the use of apron buses and the two doors of the airplane, testing classical and new methods for airplane boarding within various conditions, proposing new methods for accelerating the boarding process, analyzing the efficiency of the boarding methods from the perspective of the passenger health [8,13,14].

In the area of sea transportation, a series of ABM approaches have been provided in the following research directions: exploring varied wind-assisted ship propulsion technologies to enhance maritime efficiency and sustainability [15], simulating nautical services within ports, thereby facilitating a comprehensive comprehension of their intricate dynamics [16], modeling the disruptions of maritime trade chokepoints and the global liquefied natural gas [17], modeling the maritime traffic in piracy-affected waters [18], exploring the effects of policy instruments on the transition of the maritime fuel system away from heavy fuel oil [19]. Furthermore, due to the spatial characteristics offered by the ABMs, they have been devised to address the intricate dynamics of evacuation scenarios in cruise ship evacuation [20].

As for the transport operating by road, a series of ABM applications have focused on presenting a robust model for tackling cross-docking challenges, thereby streamlining logistics operations [21], facilitating simulations to gauge the ramifications of introducing shared bikes and e-scooters on existing travel modalities [22], scrutinizing the effects of automated mobility-on-demand services on the broader public transportation landscape [23], experimenting with the significance of traffic information exchange in ameliorating traffic congestion [24], conducting a comparative analysis between two operational strategies of public transport services, namely fixed-route transit (FRT) and demand-responsive transport (DRT) [25], employing simulations to bolster urban public transport systems, with direct implications for contingency planning [26], attaining cooperation in road networks [27], simulating the diffusion of information between road freight transport agents [28]. Another area of research connected to road transportation is related to the risks associated with this type of transportation. As a result, Rad et al. [29] study the behavior of the pedestrian while road crossing, especially in the context of the occurrence of the automated vehicles, while Shin [30] proposes an ABM for quantifying the health effects of exposure to non-exhaust road emissions to both drivers and pedestrian groups. Even in the case of road transportation, evacuation issues have been studied in the scientific literature, mostly related to the analysis of the traffic evacuation in road networks [31].

For the transport operating by rail, especially by train or metro, a series of applications have focused on unveiling the intricate interplay between individual behaviors and the overarching performance of high-speed train systems [32] and delving into the influence of behavioral parameters on modal shifts within the realm of public transportation [33]. Furthermore, the inherent advantages offered by agent-based modeling have been harnessed in the development of models pertaining to passenger evacuation from diverse modes of rail transportation: metro station evacuation [34], large transit terminal subway station [35], urban rail transit evacuation [36].

In the broader domain of transport planning, Kagho et al. [37] have astutely observed an escalating necessity for the development of ABMs that can efficaciously facilitate decision-making processes by offering a nuanced comprehension of traveler behavior. This necessitates a comprehensive understanding of intricate travel patterns, coupled with a keen awareness of their inherent sensitivity to the dynamics of the transport system.

Further studies are needed to explore the role and the conceptual frameworks of ABMs in various fields of transportation research, along with the advancements in agent-based modeling and/or similar modeling techniques in transportation.

Hybrid approaches in transportation modeling could also represent a possible research direction.

Nevertheless, specific applications in transportation solvable using ABM could be approached, especially in the context of the recent events generated by the occurrence of the COVID-19 pandemic. As it has been noted in the scientific literature, the introduction of social distancing measures within the transportation sector has notably curtailed the operational capacities of various modes of conveyance, encompassing trains, metros, apron buses, and airplanes. This impact is evident in the documentation from COBUS Industries GmbH, wherein the authors elucidate the repercussions of adhering to social distancing protocols in the COBUS models used in transportation. As an illustrative instance, the COBUS 2700s model experiences a substantial reduction as a result of the social distance, accommodating merely 10 passengers as opposed to its previous capacity of 77 passengers. Similarly, the COBUS 3000 model’s carrying capacity dwindles from 110 passengers to a mere 17 individuals, a direct outcome of conforming to social distancing regulations [38]. Through the COVID-19 pandemic, numerous other transportation service providers have undertaken comparable adaptations by curbing the seating capacity of their vehicles in alignment with the stipulations of social distancing directives. This approach has been substantiated by Gkiotsalitis and Cats [39], who cite the pertinent literature to underscore cases where city buses were constrained to ferry a maximum of 15 passengers, while rail cars could accommodate no more than 30 passengers, all within the ambit of social distancing measures. Consequently, these limitations precipitate a marked reduction, permitting a mere 10% of the usual passenger volume to take part in these journeys. The authors’ investigation into public transport ridership during the COVID-19 pandemic furnishes compelling evidence, wherein the implementation of a 1.5-m social distance policy results in the Washington DC metro system operating at a mere 18% of its total capacity. The imposition of a more stringent 2-m social distance policy further exacerbates this capacity contraction, reducing it to a mere 10% [26]. Given the profound diminution in capacity witnessed in single-destination public transportation contexts, Moore et al. [40] proffer a proactive solution by advancing a solution to address passenger seat assignments. This innovative model and others particularly adapted for special situations (such as the occurrence of a pandemic in the future) can be modeled through the use of ABMs, offering supplementary details for tailoring the capacity of transit vehicles and for engendering a more efficient utilization of the available spatial resources.