Skip to content

ccam.md

Latest commit

 

History

History
36 lines (22 loc) · 2.79 KB

ccam.md

File metadata and controls

36 lines (22 loc) · 2.79 KB

CCAM simulations for Brussels, Madrid, and Paris

The following sections describe the necessary steps to run the CCAM simulations for the Brussels, Madrid, and Paris use cases.

Prerequisites

Network data

To perform the simulation runs, network data for the individual use cases needs to be made available to the processing pipeline. For that, have a look at the guide for generating the demand data and obtain the OpenStreetMap (OSM) data referenced there. Place the respective data sets into ccam/resources/{case}/osm with case being any of paris,madrid,brussels.

Running the simulations

After collecting the files, navigate into the ccam folder using your command line and execute the following command:

poetry run snakemake -c1 -s workflow/paris.smk all

This will run the whole simulation pipeline for the Paris case. Change the paris name in the command line to brussels or madrid to run the other cases. The simulation will run the following steps:

  • Obtain the road network from OpenSteetMap
  • Load the demand data. Those data must be placed in demand/results/{case}/trips/passenger_trips_seed{seed}.gpkg, i.e., the demand generation pipeline must have been run already.
  • Enrich the trip data with service-specific attributes. Those are, for instance, defined in ccam/resources/paris/enrichment.yml and analogously for the other cases. They define, in particular, the characteristics of the user profiles (group size distribution, maximum wait time, ...).
  • Prepare the relevant data for performing a CCAM simulation (e.g., obtaining a travel time matrix across the network area).
  • Perform the simulations which are parameterized by (1) a random seed, (2) a fleet size, (3) a vehicle capacity, (4) and an operating area.

The fleet sizes to be run and the vehicle capacity are defined in ccam/workflow/paris.smk whereas the spatial extents are defined in ccam/resources/paris/spatial.gpkg and, analogously, for the user cases.

Note that spatial.gpkg contains multiple polygons. Each polygon has a name that indicates the operating area that the polygon refers to. Furthermore, there is always one polygon called name="network" which defines the extent of the routable network that will be extracted from OpenStreetMap data.

The result of each pipeline are MATSim runs located in ccam/results/{case}/matsim (for expert users) and aggregated analysis information (using the analysis scripts developed in MAIA) in ccam/results/{case}/analysis.

The aggregated outputs can be visualized using the ccam/Analysis.ipynb notebook which can be started using standard tools such as VSCode or Jupyter.