Skip to content

towards a working Frank-Wolfe implementation of the equilibrium assignment in Python

Notifications You must be signed in to change notification settings

megacell/python-traffic-assignment

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

python-traffic-assignment

This is a Python implementation of the equilibrium assignment using the traditional Frank-Wolfe algorithm. At its heart, it relies on a All-or-Nothing assignment module which original Cython implementation is from: http://www.xl-optim.com/python-traffic-assignment/.

Our Python implementation has been tested against Matthew Steel's C++ implementation of Bar-Gera's and Dial's origin-based solver: http://www.repsilat.com/EquilibriumSolver.html. Both yield same edge flow assignment on some of the test problems in: http://www.bgu.ac.il/~bargera/tntp/

Transportation networks have been moved to this repository: https://github.com/bstabler/TransportationNetworks

Setup

Compile the Cython code with the following command:

python setup_Assignment.py build_ext --inplace

Test the compiled code with:

python -m unittest discover

Running on networks from Bar-Gera's test problems

For example, the Chicago sketch network (387 zones; 933 nodes; 2950 links) can be found in: http://www.bgu.ac.il/~bargera/tntp/. Download 'ChicagoSketch_net.txt', 'ChicagoSketch_node.txt', and 'ChicagoSketch_trips.txt' files. Convert into .csv files compatible with our algorithm:

python scripts.py

'process_chicago_network()' routine generates the right input files.

Visualization

You can visualize:

python scripts_chicago.py

'visualize_equilibrium_in_chicago()' routines generates geojson file 'visualization/links.js'. Then view the network with view_network.html.

About

towards a working Frank-Wolfe implementation of the equilibrium assignment in Python

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published