An agent-based fleet management model for first- and last-mile services

Abstract

With the growth of cars and car-sharing applications, commuters in many cities, particularly developing countries, are shifting away from public transport. These shifts have affected two key stakeholders: transit operators and first- and last-mile (FLM) services. Although most cities continue to invest heavily in bus and metro projects to make public transit attractive, ridership in these systems has often failed to reach targeted levels. FLM service providers also experience lower demand and revenues in the wake of shifts to other means of transport. Effective FLM options are required to prevent this phenomenon and make public transport attractive for commuters. One possible solution is to forge partnerships between public transport and FLM providers that offer competitive joint mobility options. Such solutions require prudent allocation of supply and optimised strategies for FLM operations and ride-sharing. To this end, we build an agent- and event-based simulation model which captures interactions between passengers and FLM services using statecharts, vehicle routing models, and other trip matching rules. An optimisation model for allocating FLM vehicles at different transit stations is proposed to reduce unserved requests. Using real-world metro transit demand data from Bengaluru, India, the effectiveness of our approach in improving FLM connectivity and quantifying the benefits of sharing trips is demonstrated.

AnyLogic specifics

For the purpose of reproducibility, this subsection provides a few implementation details that are specific to AnyLogic. These details are not critical for simulation platforms built using other tools, and hence the reader can skip this material without any loss of continuity.

The discrete-event model that triggers the occurrence of certain time-ordered processes (Borshchev and Filippov 2004) is executed in AnyLogic using its process modelling library blocks. Individual events can also have a timestamp associated with them and can be scheduled in advance. To create vehicle agents for every metro station, we use the resource pool functionality of AnyLogic’s process modelling library. The number of resource units in each pool can be specified along with other properties such as vehicle speed and initial location. An advantage of using resource pools is that it captures utilisation statistics, such as the time for which FLM vehicles are idle or busy during the simulation. These can be computed using a set of process modelling library blocks as shown in Figure 17. When an instance trip (agent type LastMileTrip or FirstMileTrip) is generated, the function ProcessTrip.take(trip) is called. ProcessTrip is an Enter block at which the FLM vehicle agents are inserted into the process modelling library blocks. TakeVehicle is a Seize block that captures units of the FLM vehicle resource pool. The control then goes to the Travelling block, which is a Delay block, until the stopDelay() function is called after an FLM vehicle serves the passenger trip request and reaches the station. After this step, the control flow goes to the ReleaseVehicle block that releases the seized FLM vehicle unit. The agents are then taken out of this process flow by the Sink block.

Fig. 17

Process modelling library blocks

For simulating first-mile passenger arrivals, the dynamic event functionality of AnyLogic was used (since it allows the creation of multiple concurrent instances, each of which is independent) and initialised with the specified parameters (such as the name of the metro station where passengers arrive). For the shareability scenarios, communication of inputs and outputs between the CVRP codes and AnyLogic is established through additional custom code using PypeLine (a Python connector library for AnyLogic).