DLR-RM
model understanding a graph
hello I try to solve an AI problem related to a graph using RL and stable baselines. but it seems like the RL model cannot understand and communicate with the graph at all and even fail in simple tasks (like just moving in the graph). I tried different methods (PPO, DQN, A2C) and different styles of observation (list of edges, adjacency matrix,...) but it seems like there is no improvement at all! it is possible at all? can you help me with this, please? thanks
Since you are using a custom environment, I ask that you fill out the custom env template, so that we can better help you.
However, I suspect that this is not really an SB3 related issue, but rather a general question about reinforcement learning applied to graph problems. So I suggest you submit your question on the RL Discord, Reddit or Stack Overflow.
