search

rdgreene / sa_tsp

Q-Learning applied to the classic Travelling Salesman Problem
18 stars 3 forks source link

Miguel work in progress #2

Closed miguelesteras closed 7 years ago

miguelesteras commented 7 years ago

1 Old files were deleted and line graph and route graphs are now optimized for parameters search.

However, when I packaged the plotting into a function I get problems. Somehow the functions from scipy are not defined. Did you encounter this problem before?

Also, I prevented stp.py code from deleting the learning variables, some of them are used for plotting.

Maybe have a look to the branch and try to run it in your machine.

Miguel :neckbeard:

miguelesteras commented 7 years ago

Meanwhile I keep working on the new :boom::boom:double-Q!!!! :boom::boom:

rdgreene commented 7 years ago

Awesome! Yes I thought that could be the case (with the deleted variables), no probs. I'm about to have dinner but I'll check the code on my machine and then approve the merge if it all runs fine!

Regarding packaging plotting into a function, have you tried importing scipy within the new function itself?