2-3.2: Optimizing Service Areas to Reduce Congestion and Enhance Equity in Access to Transportation Systems

Abstract

As new transportation alternatives emerge and existing services are redesigned, communities and their planning agencies are often required to identify transportation service areas within which the services can operate. The selection of which neighborhoods are included in or excluded from a service area can have significant implications on the level of accessibility, economic sustainability, and congestion reduction potential of the transportation system. In this project, spatial optimization methods are proposed to explicitly consider the service area selection problem in the design of transportation systems. The main contribution of this project is twofold. First, optimization models are presented for service area design problems that simultaneously account for the goals of system operators and users, with particular emphasis on the objectives of reducing congestion and enhancing equity in access to transportation services. The models will account for spatial coverage constraints that are specified in real-world situations to ensure equitable access to transportation services. The second contribution of the project will be heuristics to solve the proposed design problems. The heuristics will combine surrogate-based and shape optimization techniques with evolutionary algorithm strategies to discover optimal service area designs. The optimization models will be formulated for two design applications. In the first application, the problem of designing transit routes and the accompanying paratransit service area will be examined. This problem is of practical interest to US transit agencies given federal regulations that set minimum spatial coverage of paratransit services based on fixed transit route alignments. In the second application, the problem of defining the service area for dockless micromobility services will be considered. The dockless e-scooter system is operating in Mayagüez, Puerto Rico and the travel demand models that will be developed for this system will serve as the test case for the second application.v