The project aims to implement TSP algorithm without the use of and available python libraries
One chromosome is the path that the travelling salesman is taking (the nodes)
The population of such chromosomes is getting randomly initialized and evaluated with a function summing their path lengths. The lower the value, the better the score.
Then in the loop population members get crossovered with each other with some probability (OX crossover mechanism) and reevaluated. If their new score is better than the last they are getting replaced with the new chromosome. Parents are picked randomly with probability reflecting their score. Parents with better scores are more likely to get picked.
If newly created chromosome from two parents has a worse score than the parent, it is mutated (again it is randomly selected if the mutation occurs) and replaces the parent. This allows for better scores but needs some parameter optimization to achieve the best score.
The best chromosome with its score gets saved separately to be displayed at the end with matplotlib.
pip install tsplib95python tsp.py <file.tsp>