Flexible Agent-based Modeling Framework to Evaluate Integrated Microtransit and Fixed-route Transit Designs: Mode Choice, Supernetworks, and Fleet Simulation

Abstract

The integration of traditional fixed-route transit (FRT) and more flexible microtransit has been touted as a means of improving mobility and access to opportunity, increasing transit ridership, and promoting environmental sustainability. To help evaluate integrated FRT and microtransit public transit (PT) system (henceforth ``integrated fixed-flex PT system'') designs, we propose a high-fidelity modeling framework that provides reliable estimates for a wide range of (i) performance metrics and (ii) integrated fixed-flex PT system designs. We formulate the mode choice equilibrium problem as a fixed-point problem wherein microtransit demand is a function of microtransit performance, and microtransit performance depends on microtransit demand. We propose a detailed agent-based simulation modeling framework that includes (i) a binary logit mode choice model (private auto vs. transit), (ii) a supernetwork-based model and pathfinding algorithm for multi-modal transit path choice where the supernetwork includes pedestrian, FRT, and microtransit layers, (iii) a detailed mobility-on-demand fleet simulator called FleetPy to model the supply-demand dynamics of the microtransit service. In this paper, we illustrate the capabilities of the modeling framework by analyzing integrated fixed-flex PT system designs that vary the following design parameters: FRT frequencies and microtransit fleet size, service region structure, virtual stop coverage, and operating hours. We include case studies in downtown San Diego and Lemon Grove, California. The computational results show that the proposed modeling framework converges to a mode choice equilibrium. Moreover, the scenario results imply that introducing a new microtransit service decreases FRT ridership and requires additional subsidies, but it significantly increases job accessibility and slightly reduces total VMT.