# cityflow-project/CityFlow

A Multi-Agent Reinforcement Learning Environment for Large Scale City Traffic Scenario

Repository: https://github.com/cityflow-project/CityFlow
Canonical: https://ross.abutalabs.com/products/cityflow
Homepage: https://cityflow-project.github.io
Language: C++
License: Apache-2.0
License Family: permissive
Topics: multiagent-systems, multiagent-reinforcement-learning, traffic-simulation, traffic-signal-control
Last push: 2025-08-19T21:39:36+00:00

## Health v2 (maintenance only)
Score: 49/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 37, release rhythm 35, longevity 100
- inputs: {"age_days": 2745, "days_push": 379, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1011, forks 205 (observed 2026-08-28T04:03:13.134578+00:00)

## What it is
CityFlow is an open-source microscopic traffic simulator designed as a multi-agent reinforcement learning environment for large-scale city traffic scenarios. It provides a fast, multithreaded C++ engine with a Python interface, significantly outperforming SUMO for city-wide traffic signal control research.

## Use cases
- train reinforcement learning agents for traffic signal control
- simulate city-wide traffic with thousands of vehicles
- benchmark multi-agent RL algorithms on road networks
- compare traffic control policies against SUMO
- build custom road networks and traffic flows for transportation research
- run reproducible traffic simulations for ML experiments

## When to choose
- you need a fast RL environment for large-scale traffic signal control
- SUMO is too slow for your city-scale simulation experiments
- you want a Python-friendly simulator for multi-agent reinforcement learning research
- reproducibility of traffic simulations matters for your work

## When to avoid
- you need detailed driver behavior modeling or mobility features beyond traffic flow
- you require SUMO's ecosystem of tools, formats, and plugins
- you need a GUI-first traffic planning tool rather than a programmatic RL environment
- your scenario is small-scale where simulation speed is not a concern

## Facets
- artifact type: library
- maturity: active
- function: simulation, machine-learning, reinforcement-learning, agent-framework
- domain: simulation, reinforcement-learning, autonomous-vehicles, machine-learning
- platform: python, cpp
- tags: traffic-simulation, traffic-signal-control, multi-agent-reinforcement-learning, gym-environment, transportation, linux, macos, docker

## Member repositories
- cityflow-project/CityFlow (main) score 49

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:03:13.134578+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T07:11:49.654202+00:00, confidence not recorded.
  - readme: https://github.com/cityflow-project/CityFlow (fetched 2026-08-28T04:03:13.134578+00:00, sha bdca7003fd7a)
  - homepage: https://cityflow-project.github.io (fetched 2026-08-29T13:11:28.582820+00:00, sha f1f7ef62f353)
- Data as of 2026-08-30T08:39:29.467469+00:00.