reroutingStudies

My implementations of centralized rerouting strategies using Python and SUMO.

Rerouting strategies

The article "Proactive Vehicle Re-routing Strategies for Congestion Avoidance"(2012) describes three different aproachs related to vehicle rerouting: DSP, RkSP and EBkSP. Those strategies were implemented varying the math model used to predict road weight. Below, all codes directories are listed:

• Initial solutions(not updated)
• Linear models
• Logarithm models

The script runAll.py it's used to run all models with previously setted parameters. To use the script it's necessary pass an integer number $n$ as a paramater indicating the number of simulations (the SUMO random seed ranges from $0$ to $n$). Below is an example of the function's call to run 5 simulations:

runAll.py 5

Other Scripts

• genCharts.py(python3 script): Do bootstrap on data and generates differents images about the outputs. Obs.: Uses pandas append (deprecated), in the future (i hope) it will be modified to pandas concat.

• runSome.py: Select the especific models to run in one simulation using EBkSP. Has the same function call as runAll.py.

Packages and scenarios

• Sumo 0.30 is available in this link. To install sumo run ./configure then make install in sumo dir. Problems related with the HUGE constant in some src files can be solved changing it to DBL_MAX (check compiler extension for maximum double variable value). Sumo also requires xerces and foxtoolkit.

• Cologne scenario in NewCologne.zip file can be directly accessed here.