1. Home
  2. RIP
  3. View Record
https://nap.nationalacademies.org/catalog/27432/critical-issues-in-transportation-for-2024-and-beyond

MMOS Integration

This project will integrate the multi-modal optimization model developed in M1 with the multi-agent simulation developed in M2. The integration in the proposed design allows each model to rely on the strengths of the other. Optimization models are prescriptive, meaning that they tell us what to do given objectives and constraints. In contrast, simulation models are descriptive, meaning that they tell us how travelers would respond to specified transit service provisions. The major effort and innovation in this project is in managing the consistency between an simulation and the optimization model. The project involves developing and testing a mechanism to enable feedback of the newly designed on-demand transit system with the agent-based simulation. This is distinct from a fixed-route transit system because the on-demand shuttles respond to the demand, while the demand responds to the availability of shuttles, making the two endogenous. Therefore, the demand simulation will consider the expected supply, then then the supply will respond to whether or not the demand agent is expected to use the shuttle. Together the integrated MMOS provides a recommended transit service provision and an estimate of the resulting ridership and impacts, summarized to generate performance metrics.  The models are being simultaneously developed in a collaboration with staff from SFCTA and MTC in San Francisco, and with WFRC and UTA in Salt Lake City. These external collaborators join weekly meetings to inform of local situations, supply data, and learn of the project’s progress and purposes.