Nash Flow Computation

Jan 28, 2022
Code

Project Objective

The Nash Flow Computation Tool (NFC) is a Python-based project that uses the SCIP mixed integer program solver to compute (spillback) thin flows, and consequently, Nash flows over time with or without spillback, as it is described in my thesis.

The network can be created by drag and drop.
The network can be created by drag and drop.

Features

NFC is embedded in a user-friendly graphical interface, enabling quick and easy creation of networks. It allows the analysis of earliest arrival time functions and inflow functions in a dynamic equilibrium, which can be computed with a single click for further research. The flow models for the case without spillback are based on the publication of Koch and Skutella (2011) and for the case with spillback on this paper by Vargas Koch and me (2019).

The application provides all relevant information about the computed Nash flow over time.
The application provides all relevant information about the computed Nash flow over time.

Tools and Libraries

The project makes use of the following tools and libraries:

  • Python 3.7
  • PyQt5
  • matplotlib
  • numpy
  • networkx
  • SCIP

Development

NFC was initially developed as part of the bachelor thesis of Max Zimmer and was later enhanced within the TU Berlin/ECMath project titled “Dynamic Models and Algorithms for Equilibria in Traffic Networks”.

My Role

As a supervisor, I guided the developer throughout the project, ensuring the successful implementation of the sophisticated Nash flow over time algorithms.

Major Optimization