Peer Review Lab 2 - Leonor Gomes

[leonormgomes]

Here is my review:

• Good results and quick execution time. It's positive that you managed to test with N = 5000 and N = 10000;
• The last cell of the notebook is quite big, I suggest you refactor it;
• It's good that you have tried two methods for crossover. Have you considered using both and choosing which one to use according to a probability?
• I like the fact that you have implemented a way to switch from exploration to exploitation. However would it be necessary to have a ARTIFICIAL_MUTATION = 50000 for small N like 5 and 10? Maybe an idea would be to have this variable related to the size of the problem;
• I would suggest a better commenting coverage, especially for some variables (Ex: it's a bit hard to quickly understand the purpose of g_a) and trickier functions/lines of code (Ex: artificial_mutation(), you explained why you use it, however it would be appreciated if you explained what you were doing as well so the difference between the first and this mutation is clearer).

[FlavioPatti]

Thank you for your review! About the crossover I didn't use probability to choose one, I simply tested the one_cut_cross_over first and then the uniform version and I realized that the results are improved so I only used the last one at the end. I agree on the 4th point, effectively for small value of N it doesn't make much sense to use large ARTIFICIAL_MUTATION value, I did not think about it ahah