This project is originated from a university course project.
In 2016, a friend of mine, majoring in logistic engineering, came to me to discuss his course work project. He wanted to solve VRPTW using genetic algorithm, which I happened to know. The discussion went well and my friend got what he needed.
After that, I implemented the approach in Python. The first version of the this project came out that night.
I wrote this project on the spur of that moment. However, after running a few tests with several combinations of parameters and set-ups, I realized that implementing the idea is one thing, and tuning the algorithm to yield a converged result is another thing. The latter would definitely require much more effort.
Therefore, I should say, to be precise, the performances of the given examples are poor.
Of course, tuning improvements of this algorithm and forks are always welcome.