pigna90 / pctsptw

Prize-Collecting Traveling Salesman Problem with Time Windows

Prize-Collecting Traveling Salesman Problem with Time Windows (PCTSPTW)

This project has been made for educational purposes during the lecture of Model-Driven Decision Methods a.a 2015/16 (University Of Pisa). The goal is to implement a Linear Programming Model with PuLP and solve it with at least two different solvers: CBC and GUROBI (Academic License) (feel free to use whatever you want/need!). At last optimize one of these in order to improve time performance.

I wrote a simple report (in Italian) as a description of the work I've done. If you need it please contact me.

Requirements

• Python 3.x
• PuLP - as Python module
• Pandas - as Python module
• Numpy - as Python module
• CBC
• GUROBI

Files and directories description

A brief description of project contents:

• module. py - implementation of PCTSPTW problem and solver call
• instance_utility. py - instances generators and other utility functions
• data/ - contains all instances
• src/ - source files

Examples usage

Solve one instance using CBC with max_time=600s:

$ python3 src/model.py instance.dat CBC 600

Solve a set of instances using GUROBI with max_time=10s:

$ python3 src/model.py /path/to/instances_folder/ GUROBI 10