product mdp

[LuryN]

Hello, I found that the implementation of agent1 and agent2 in your code is almost the same, I can't find the code part about Modular Deep RL, where is the code part that combines mdp and ltl?

[spideralessio]

Hey, agent 1 and 2 share the same architecture but are two different instances (different weights). The ltl model is in the slides linked in the readme.

Basically our ltl has 2 steps, the first task is to reach the first target and then the second target. This is actually translated in a simple if case in our code. This was just a simple example for an exam project.