HOWEVER, in the paper "Multi-Agent Active Search: a Reinforcement Learning Approach", it's said that it used 2 scalars (mean and log-variance) for the altitude and 2 scalars (mean and log-variance) for the x-coordinate.
Is there a change/conflict/inconsistence?
Thanks a lot.
In the environment, such as AgnosticMAAS: https://github.com/cvigoe/DRL4MAAS/blob/95539197c9b82a34f9128fd265749d0f8f76157f/scripts/variant.py#L51 It seem that action is the segment (left and right boundary) an agent is going to sense, according to: https://github.com/cvigoe/DRL4MAAS/blob/95539197c9b82a34f9128fd265749d0f8f76157f/gym_activesearchrlpoisson/gym_activesearchrlpoisson/envs/agnosticmaas_env.py#L258-L265
HOWEVER, in the paper "Multi-Agent Active Search: a Reinforcement Learning Approach", it's said that it used 2 scalars (mean and log-variance) for the altitude and 2 scalars (mean and log-variance) for the x-coordinate.
Is there a change/conflict/inconsistence? Thanks a lot.