Python implementation of MATSim

Plan

Python interfaces for experimenting with MATSim flavoured simulations.

• Parallel sims
• Parallelised sims
• ML planning

WIP

• API for (i) planners and (ii) simulations.
• Data design for (i) simulation instructions, and (ii) simulation events.
• Tooling for (i) initialisation e.g. loading scenarios, and (ii) reporting e.g. progress and utilization.
• Planner is responsible for making instructions:
  • Initial case
  • Based on prior events from one or many sims
• Sim is responsible for resolving instructions:
  • Stay - duration for an activity
  • Move - move from entity to entity
• Stay for fixed duration:
  • Facilities for activity
  • (depending on capacity e.g. car park or building or charging)
• Move from A to B:
  • Facility to Link
  • Link to Link
  • (Stop to Stop)
• Loading:
  • Plans
  • Networks
  • (Facilities)
  • (Transit)
  • (Vehicles)
• Reporting:
  • trip durations
  • trip distances
  • link speeds
  • (utility)

Installation

Currently using uv.

So to run a demo notebook:

git clone git@github.com:fredshone/mobslim.git
cd mobslim

# then take a look at the demos

# if you need, spin up a notebook as follows:
uv run --with jupyter jupyter lab