simom0/tsp

TSP optimization, Operations Research 2 project, UniPD 2022/23

Travelling Salesman Problem (TSP) Optimization

Repository for the Operations Research 2 project regarding the TSP optimization, UniPD 2022/23.

The goal of the project is to solve the Travelling Salesman Problem with different approaches such as constructive heuristics, refinement heuristics, metaheuristics, exact methods, matheuristics ...

Project Structure

TSP
├── CMakeLists.txt          Project CMakeLists
├── README.md               Project README
├── build                   Generated by CMake
│   ├── ...
│   ├── main                Main executable file
├── data                    TSP data file
│   └── att48.tsp           Example .tsp file
├── include                 Header files
│   ├── ...
│   ├── tsp.h
│   └── utils.h
├── plot                    Plot files
│   └── att48.png           Example plot file
└── src                     Source files
    ├── ...
    ├── main.c
    └── utils.c

Requirements

• Gnuplot: a portable command-line driven graphing utility
• CPLEX: IBM CPLEX Optimizer
• Concorde: TSP solver

Installation

• Create the build directory using mkdir build
• Navigate into the build directory with cd build
• Run the command cmake .. to configure the project
• Execute the make command to build the project

Usage

Input data are in .tsp from TSPLIB. Some of them are already stored in the data directory.

Run the main program with the following command and set the desired parameters:

./main -file <filepath> -solver <solver> -time_limit <tl> -seed <s> -verbose <v>

Example: solve the TSP regarding the att48 instance stored in the ../data/att48.tsp file with the greedy algorithm, running the command:

./main -file ../data/att48.tsp -solver GREEDY -time_limit 60 -seed 10 -verbose 10

To list all the parameters and solvers available run ./main -help or ./main --help

To create a random testbed run ./main -testbed <num_instances> <num_points> where num_instances is the number of TSP instances to be generated and num_points the size of each instance.