RTSLearn is an experimental project to re-learn C++.
The goal is to create a performant ASCII-based RTS for use in deep learning environments similar to OpenAI's gym.
[WIP] 1 episode of 1,000,000 steps with fog of war, sampling random actions.
10x10 map:
20x20 map:
50x50 map:
Tested on i7-6700K @ 4GHz