Scenarios¶
1. Pick the stage¶
Short-run responses (route and departure time) need only the assignment. Responses in destination and mode need MAST with new skims; long-run responses in where people live and work need the land-use model too.
| Response | Stages to rerun |
|---|---|
| Route, departure time | MATSim |
| Destination, mode, trip making | MAST → MATSim |
| Residential and workplace location, car ownership | SILO → schools → MAST → MATSim |
2. A pricing scenario in MATSim¶
- Write a road-pricing file (MATSim
roadpricingformat): the links, the time windows and the amount. The base fileassignment/network/roadpricing_HARB_2024.xmlis an example. -
Run the base and the scenario with the same plans and seeds:
-
Compare executed trips (
output_trips.csv.gz), link volumes (link_hours_s<N>.csv.gz) and scores (output_plans.xml.gz) between base and scenario, seed by seed.
Seeds
Run at least five seeds for base and scenario. Differences smaller than the spread across seeds are noise.
3. Reading results by person¶
Every MATSim agent carries its household income, cars, home zone and county, so changes in travel time, toll paid and utility can be summarised by income decile or by area.
4. Land-use scenarios¶
Edit the land-use inputs (development_*.csv for building capacity, job control totals for employment), write a
properties file with landuse/run/write_properties.py, and run SILO from 2016; then rerun schools, MAST and
MATSim on the scenario's 2024 population.