Solving the Traveling Salesman Problem using Self-Organizing Maps

• Changes from Y-Shy@Github：
  • A ch34.tsp data file was added: 34 cities from china. Hong Kong, Macao, Taipei, Haikou included
  • Changes in ./src/main.py
    • Argv related lines were deleted
    • File pathes were modified
  • Changes in ./src/plot.py Linewidth and marksize were enlarged so that you would get a clearer figure if you wanted to print it.

This repository contains an implementation of a Self Organizing Map that can be used to find sub-optimal solutions for the Traveling Salesman Problem. The instances of the problems that the program supports are .tsp files, which is a widespread format in this problem. All the source code can be found in the src directory, while a report and brief presentation slides (in Spanish) can be found in the report folder. However, for a complete read on the topic, you can read my blog post explaining this implementation and its evaluation.

To run the code, only Python 3 and the dependencies (matplotlib, numpy and pandas, which are included in the Anaconda distribution by default) are needed. In case you are not using Anaconda, you can install all the dependencies with:

pip install -r requirements.txt

To run the code, simply execute:

cd som-tsp
python src/main.py assets/<instance>.tsp

The images generated will be stored in the diagrams folder. Using a tool like convert, you can easily generate an animation like the one in this file by running:

convert -delay 10 -loop 0 *.png animation.gif

This code is licensed under MIT License, so feel free to modify and/or use it in your projects. If you have any doubts, feel free to contact me or contribute to this repository by creating an issue.

---

This code was presented for the Bio-Inspired Artificial Intelligence course in the Computer Science & Technology master’s degree @ UC3M. A previous version of this code can be found in this repository. Special thanks to Leonard Kleinans, who worked with me in that previous version.