I tried to adapt the environment to multiple contexts. I tried it with sample values and it seems to work correctly. Let me know what you think.
Basically the main idea is to allocate the budget to different classes not in a uniform way but with a criteria based on the output of the contextual learner. The rest of the code is the same.
I tried to adapt the environment to multiple contexts. I tried it with sample values and it seems to work correctly. Let me know what you think. Basically the main idea is to allocate the budget to different classes not in a uniform way but with a criteria based on the output of the contextual learner. The rest of the code is the same.