Abstract
Integrating microtransit with fixed-route transit (FRT) can improve travelers’ mobility by leveraging the benefits of microtransit’s flexibility and FRT’s high passenger capacity. However, the high operating cost of microtransit presents a challenge, which calls for a careful evaluation of the trade-offs between mobility gains and operational cost. To address this need, this paper develops a modeling approach and solution procedure to identify Pareto-optimal designs. We focus on design parameters of practical interest, namely, fare policies and microtransit fleet size. To explore these trade-offs, we propose a bi-level and bi-objective (i.e., minimize taxpayer subsidy and maximize mobility-based consumer welfare) modeling framework, with an agent-based transportation system simulation model at the lower level and a multi-objective Bayesian Optimization (BO) model at the upper level. We apply the modeling and solution approach to Lemon Grove, California (a suburban area in San Diego County). Results reveal a diverse set of solutions along the Pareto frontier, indicating that some naive microtransit fare strategies are suboptimal. Notably, Pareto-optimal designs feature a 50-100% discount for microtransit to FRT transfers, as well as peak-period fare multipliers between 1.8x and 3.5x to manage time-varying demand effectively.