UXsim is a free, open-source macroscopic and mesoscopic network traffic flow simulator written in Python. It simulates the movements of car travelers and traffic congestion in road networks. It is suitable for simulating large-scale (e.g., city-scale) traffic phenomena. UXsim is especially useful for scientific and educational purposes because of its simple, lightweight, and customizable features, but users are free to use UXsim for any purpose.
If you are interested, please see:
- Jupyter Notebook or Google Colab: Interactive demonstrations and tutorials
- Technical Documentation: Detailed documents on tutorials, simulation mechanism, and specifications of modules/functions
- Simple, lightweight, and easy-to-use Python implementation of modern standard models of dynamic network traffic flow
- Macroscopic traffic simulation: Simulating over 60000 vehicles in a city in 30 seconds, or even 1 million vehicles in a metropolitan area in 40 seconds depending on the simulation setting
- Dynamic traffic assignment: Traffic flow simulation with a given network and time-dependent OD demand
- Theoretically valid models commonly used in academic/professional transportation research
- Implementation of traffic control/management schemes such as taxi/shared-mobility, traffic signals, road pricing, and so on
- Basic analysis of simulation results and their export to
pandas.DataFrameand CSV files - Visualization of simulation results using
Matplotlib; interactive GUI is also available - Flexible and customizable thanks to pure Python implementation; can also be directly integrated with other Python-based frameworks, such as
PyTorchfor deep reinforcement learning traffic control - The main code
uxsim.pyis only about 1800 lines of code. Users may easily understand and customize it
Below are simulation results where approximately 60000 vehicles pass through a 10km x 10km grid network in 2 hours. The computation time was about 30 seconds on a standard desktop PC.
Visualization of link traffic states (thicker lines mean more vehicles, darker colors mean slower speeds) and some vehicle trajectories:
Vehicle trajectory diagram on a corridor of the above network:
A traffic signal controller is trained by deep reinforcement learning (DRL) using PyTorch. The left (or upper) scenario shows no control with fixed signal timing; the traffic demand exceeds the network capacity with the naive signal setting, and a gridlock occurs. The right (or bottom) scenario shows DRL control, where the traffic signal can be changed by observing queue length; although the demand level is the same, traffic flows smoothly. A Jupyter Notebook of this example is available.
resultGUIviewer_cut.mp4
The simplest way is to use pip to install from PyPI:
pip install uxsim
You can also install with conda from conda-forge channel:
conda install uxsim
For the details, please see here.
Alternative methods for advanced users (click to see)
You can also use pip to install the GitHub version:
pip install -U -e git+https://github.com/toruseo/uxsim@main#egg=uxsim
Or any other (development) branch on this repo or your own fork:
pip install -U -e git+https://github.com/YOUR_FORK/uxsim@YOUR_BRANCH#egg=uxsim
Download the uxsim directory from this Github repo or the latest release and place it in your local directory as follows:
your_project_directory/
├── uxsim/ # The uxsim directory
│ ├── uxsim.py # The main code of UXsim. You can customize this as you wish
│ └── ... # Other files and directories in uxsim
├── your_simulation_code.py # Your code if necessary
├── your_simulation_notebook.ipynb # Your Jupyter notebook if necessary
├── ... # Other files if necessary
This way, you can flexibly customize UXsim on your own.
As a simple example, the following code will simulate traffic flow in a Y-shaped network.
from uxsim import *
# Define the main simulation
# Units are standardized to seconds (s) and meters (m)
W = World(
name="", # Scenario name
deltan=5, # Simulation aggregation unit delta n
tmax=1200, # Total simulation time (s)
print_mode=1, save_mode=1, show_mode=0, # Various options
random_seed=0 # Set the random seed
)
# Define the scenario
## Create nodes
W.addNode(name="orig1", x=0, y=0)
W.addNode("orig2", 0, 2)
W.addNode("merge", 1, 1)
W.addNode("dest", 2, 1)
## Create links between nodes
W.addLink(name="link1", start_node="orig1", end_node="merge",
length=1000, free_flow_speed=20, number_of_lanes=1)
W.addLink("link2", "orig2", "merge", length=1000, free_flow_speed=20, number_of_lanes=1)
W.addLink("link3", "merge", "dest", length=1000, free_flow_speed=20, number_of_lanes=1)
## Create OD traffic demand between nodes
W.adddemand(orig="orig1", dest="dest", t_start=0, t_end=1000, flow=0.45)
W.adddemand("orig2", "dest", 400, 1000, 0.6)
# Run the simulation to the end
W.exec_simulation()
# Print summary of simulation result
W.analyzer.print_simple_stats()
# Visualize snapshots of network traffic state for several timesteps
W.analyzer.network(100, detailed=1, network_font_size=12)
W.analyzer.network(600, detailed=1, network_font_size=12)
W.analyzer.network(800, detailed=1, network_font_size=12)It will output text to the terminal and images to the out directory like below:
simulation setting:
scenario name:
simulation duration: 1200 s
number of vehicles: 810 veh
total road length: 3000 m
time discret. width: 5 s
platoon size: 5 veh
number of timesteps: 240
number of platoons: 162
number of links: 3
number of nodes: 4
setup time: 0.00 s
simulating...
time| # of vehicles| ave speed| computation time
0 s| 0 vehs| 0.0 m/s| 0.00 s
600 s| 130 vehs| 13.7 m/s| 0.03 s
1195 s| 75 vehs| 12.3 m/s| 0.06 s
simulation finished
results:
average speed: 11.6 m/s
number of completed trips: 735 / 810
average travel time of trips: 162.6 s
average delay of trips: 62.6 s
delay ratio: 0.385
To learn more about UXsim, please see:
- Simple demo in Jupyter Notebook or Google Colab: Interactive demonstrations
- Demos and examples: Various examples using Jupyter Notebooks and Python codes
- UXsim Technical Documentation: Detailed documents on tutorials, simulation mechanism, and specifications of modules/functions
- arXiv preprint: Scientific overview
uxsimdirectory: UXsim main packageuxsim/uxsim.py: UXsim main code
demos_and_examplesdirectory: Tutorials and examples of UXsimdatdirectory: Sample scenario files
UXsim is released under the MIT License. You are free to use it as long as the source is acknowledged.
When publishing works based on UXsim, please cite:
- Toru Seo. Macroscopic Traffic Flow Simulation: Fundamental Mathematical Theory and Python Implementation. Corona Publishing Co., Ltd., 2023.
- Toru Seo. UXsim: An open source macroscopic and mesoscopic traffic simulator in Python-a technical overview. arXiv preprint arXiv: 2309.17114, 2023
Works using UXsim is summarized on the Github Wiki page. Please feel free to edit.
Contributions are welcome! Please see the Contributing Guideline.
If you have any questions or suggestions, please post them to the Issues or Discussions (in English or Japanese).
I (Toru Seo) work on this project in my spare time. Please understand that my response may be delayed.
UXsim is based on various works in traffic flow theory and related fields. We acknowledge the contributions of the research community in advancing this field. Specifically, UXsim directly uses the following works:
- Newell's simplified car-following model and its extension X-model
- Incremental Node Model and its mesoscopic version
- Dynamic User Optimum-type Route Choice Model
![@dependabot[bot]](https://avatars.githubusercontent.com/in/29110?s=64&v=4){kind=small}