Skip to content

Latest commit

 

History

History
53 lines (41 loc) · 2.93 KB

README.md

File metadata and controls

53 lines (41 loc) · 2.93 KB
Raw

T-Score Scenarios

The MMOS half of the T-SCORE UTC revolves around applying and adapting a multi-agent simulation and an multi-modal optimization algorithm to examine the potential consequences of ride-hail and on-demand transit policies in multiple cities. This repository contains the data necessary to run these scenarios, as well as provide a template for applications in other places.

The MMOS scenarios are developed for two metropolitan regions:

  • San Francisco County, California
  • Wasatch Front region (Salt Lake City and environs), Utah

Project Navigation

There are four key folders in this repository:

  1. convert: Contains the files created externally, e.g. MATSim network file, ride-hail services etc, that are to be converted to BEAM supported format.
  2. output: Contains the total output generated from a BEAM run. Follows the convention specified in BEAM user guide. Will be locally available only, i.e. ignored while committing to github.
  3. sf-tscore: Contains the relevant input and output files for MMOS scenario in San Francisco County, California.
  4. slc-tscore: Contains the relevant input and output files for MMOS scenario in Wasatch Front region (Salt Lake City and environs), Utah.

The How to Run TSCORE Scenario markdown file describes the input files necessary for scenario analyses and how to run these scenario.

Handoff Strategy

This section describes the specifics of how the two different halves of the MMOS project interact. This was originally described in the proposal with a version of the following graphic:

MMOS Diagram

The details are as follows:

  1. Scenario-specific plans and network files are developed as inputs to the BEAM multi-agent transportation simulation.
  2. The BEAM mode choice model and transit / ride hail dispatcher provide an estimate of which individuals are likely to use each service, adapted from the initial choice supplied by the activity-based models.
    • An initial guess as to the location and fleet size of ridehail and microtransit vehicles will be necessary.
    • On-demand microtransit will be represented as a pooled ridehail option. This means that the existing BEAM ridehail manager may need modest changes to represent lower fares or more restrictive geofencing.
  3. BEAM output files for highway and transit networks are provided as inputs to the multi-modal optimization program (SAM?). Also supplied is the the BEAM output plans file, with which mode each individual used for each trip.
  4. SAM will generate a new fixed-route transit network and a list of ridehail and on-demand transit requests.
    • We will need to develop a strategy to convert these request frequencies and locations into a fleet design, which will feed back into step 2.

Both halves of the model will use the same mode choice model utility functions.