Transit Network Optimisation

This repository provides a framework for the joint optimization of public transport schedules and Demand Responsive Transit (DRT) fleet sizes.

For an overview of the research area, see Planning, operation, and control of bus transport systems: A literature review. The specific focus of this implementation is the frequency setting problem, extending it to multimodal networks (PT frequency setting + DRT discrete fleet sizing). A relevant foundational paper is Joint design of multimodal transit networks and shared autonomous mobility fleets.

Features

• Encoding
  • Parse and extract operational data directly from raw GTFS feeds.
• Optimisation
  • Headway optimization of existing transit networks using metaheuristics (PSO via PyMOO).
  • Joint optimization of Fixed-Route PT headways and DRT fleet sizing.
  • Pluggable framework for customized objective functions (e.g., waiting time, service coverage) and configurable constraints (e.g., fleet limits).
• Decoding
  • Automated reconstruction of optimized solutions back into valid GTFS outputs for downstream simulation (e.g., MATSim).

Installation

This project uses uv for dependency management. To set up the virtual environment:

uv venv --python 3.12 .transit_opt_uv
source .transit_opt_uv/bin/activate
uv pip install -e .
uv pip install -r requirements.txt

Usage

The core functional pipeline runs via the CLI using a YAML configuration file. The configuration defines the input GTFS, the allowable headway parameters, optimization constraints, evaluation intervals, and output directories.

To run the optimization pipeline:

python scripts/run.py --config configs/config_template.yaml

If you are running iterative setups (such as starting an optimization run with seeded samples from a previous run), you can specify the target iteration index:

python scripts/run.py --config configs/iteration_01/your_config.yaml --iteration 1

Notebooks

You can find conceptual walk-throughs, data visualisations, and implementation examples in the notebooks/ directory.

Warning

While notebooks are a helpful starting point to understand the framework's mechanics, some may be outdated compared to the active scripts/run.py pipeline.

Credits

This package was created with Cookiecutter and the audreyfeldroy/cookiecutter-pypackage project template.