Skip to content

Setting Up the Optimization Environment for Traffic Assignment Problem

Jump to bottom
ASU Trans+AI Lab edited this page Nov 3, 2024 · 3 revisions

Installation Guide

This guide provides step-by-step instructions for setting up the required environment for running optimization codes using taplite.

Step 1: Install Anaconda

  1. Visit Anaconda's official website and register to receive the installation link via email.
  2. Download and install Anaconda on your system.

Step 2: Install Required Packages

To set up the necessary Python libraries, follow these steps:

  1. Install Ipopt
    Go to the Ipopt package page and run:

    conda install conda-forge/label/cf201901::ipopt

  2. Install Pyomo
    Go to the Pyomo package page and run:

    conda install conda-forge::pyomo

  3. Install CVXPY
    Go to the CVXPY package page and run:

    conda install conda-forge::cvxpy

Step 3: Access Optimization Codes

  1. Locate the Python codes for the optimization model at ASU Trans-AI Lab’s GitHub repository.
  2. Review the following code files:
    • traffic_assignment_cvxpy.py
    • traffic_assignment_ipopt.py

Step 4: Run the Code in Spyder

  1. Open Anaconda Navigator.
  2. Launch the Spyder Python environment.
  3. In Spyder, open and run the selected code files (traffic_assignment_cvxpy.py or traffic_assignment_ipopt.py) to execute the traffic assignment models.

Example Output (CVXPY)

(CVXPY) Nov 03 10:05:30 AM: Problem status: optimal_inaccurate
(CVXPY) Nov 03 10:05:30 AM: Optimal value: 2.040e+04
(CVXPY) Nov 03 10:05:30 AM: Compilation took 1.689e-02 seconds
(CVXPY) Nov 03 10:05:30 AM: Solver (including time spent in interface) took 9.023e-02 seconds

Optimal Flows on Links:

  • Flow on link (1, 3): 944.1528 vehicles
  • Flow on link (3, 2): 944.1593 vehicles
  • Flow on link (1, 4): 48.7980 vehicles
  • Flow on link (4, 2): 48.7957 vehicles

Objective value (Total Cost): 20399.6443

Alternative Result on a slightly different problem (IpOPT):

  • Flow on link (1, 3): 5569.5249 vehicles
  • Flow on link (3, 2): 5569.5249 vehicles
  • Flow on link (1, 4): 2430.4751 vehicles
  • Flow on link (4, 2): 2430.4751 vehicles

Objective value (Total Cost): 197485.3688

Path Flows:

  • Flow on Path 1 (1->3->2): 5569.5249 vehicles
  • Flow on Path 2 (1->4->2): 2430.4751 vehicles

Path Travel Times:

  • Travel Time on Path 1 (1->3->2): 31.2760 minutes
  • Travel Time on Path 2 (1->4->2): 31.2760 minutes

Minimum Travel Time: 31.2760 minutes

User Equilibrium Gaps:

  • Gap for Path 1 (1->3->2): 0.0000 minutes
  • Gap for Path 2 (1->4->2): 0.0000 minutes

Weighted Gaps (by Flow):

  • Weighted Gap for Path 1 (1->3->2): 0.0000 vehicle-minutes
  • Weighted Gap for Path 2 (1->4->2): 0.0836 vehicle-minutes
  • Total Weighted Gap: 0.0836 vehicle-minutes

Objective value (Total Cost): 197485.3688

Installation Guide

This guide provides step-by-step instructions for setting up the required environment for running optimization codes using taplite.

Step 1: Install Anaconda

  1. Visit Anaconda's official website and register to receive the installation link via email.
  2. Download and install Anaconda on your system.

Step 2: Install Required Packages

To set up the necessary Python libraries, follow these steps:

  1. Install Ipopt
    Go to the Ipopt package page and run:

    conda install conda-forge/label/cf201901::ipopt

  2. Install Pyomo
    Go to the Pyomo package page and run:

    conda install conda-forge::pyomo

  3. Install CVXPY
    Go to the CVXPY package page and run:

    conda install conda-forge::cvxpy

Step 3: Access Optimization Codes

  1. Locate the Python codes for the optimization model at ASU Trans-AI Lab’s GitHub repository.
  2. Review the following code files:
    • traffic_assignment_cvxpy.py
    • traffic_assignment_ipopt.py

Step 4: Run the Code in Spyder

  1. Open Anaconda Navigator.
  2. Launch the Spyder Python environment.
  3. In Spyder, open and run the selected code files (traffic_assignment_cvxpy.py or traffic_assignment_ipopt.py) to execute the traffic assignment models.

Example Output (CVXPY)

(CVXPY) Nov 03 10:05:30 AM: Problem status: optimal_inaccurate
(CVXPY) Nov 03 10:05:30 AM: Optimal value: 2.040e+04
(CVXPY) Nov 03 10:05:30 AM: Compilation took 1.689e-02 seconds
(CVXPY) Nov 03 10:05:30 AM: Solver (including time spent in interface) took 9.023e-02 seconds

Optimal Flows on Links:

  • Flow on link (1, 3): 944.1528 vehicles
  • Flow on link (3, 2): 944.1593 vehicles
  • Flow on link (1, 4): 48.7980 vehicles
  • Flow on link (4, 2): 48.7957 vehicles

Objective value (Total Cost): 20399.6443

Alternative Result on a slightly different problem (IpOPT):

  • Flow on link (1, 3): 5569.5249 vehicles
  • Flow on link (3, 2): 5569.5249 vehicles
  • Flow on link (1, 4): 2430.4751 vehicles
  • Flow on link (4, 2): 2430.4751 vehicles

Objective value (Total Cost): 197485.3688

Path Flows:

  • Flow on Path 1 (1->3->2): 5569.5249 vehicles
  • Flow on Path 2 (1->4->2): 2430.4751 vehicles

Path Travel Times:

  • Travel Time on Path 1 (1->3->2): 31.2760 minutes
  • Travel Time on Path 2 (1->4->2): 31.2760 minutes

Minimum Travel Time: 31.2760 minutes

User Equilibrium Gaps:

  • Gap for Path 1 (1->3->2): 0.0000 minutes
  • Gap for Path 2 (1->4->2): 0.0000 minutes

Weighted Gaps (by Flow):

  • Weighted Gap for Path 1 (1->3->2): 0.0000 vehicle-minutes
  • Weighted Gap for Path 2 (1->4->2): 0.0836 vehicle-minutes
  • Total Weighted Gap: 0.0836 vehicle-minutes

Objective value (Total Cost): 197485.3688

Clone this wiki locally