How can I tune the PPO and other policy gradient family algorithms exploration-explotation behavior? which parameters should I use to fine-tune that?

Exploration in PPO and policy gradient algorithms

arturn November 21, 2021, 9:41pm 2

Give this a read!
The parameters you are looking for are entropy_coeff and entropy_coeff_schedule.

Cheers

1 Like

Topic		Replies	Views
Tuning entropy in PPO RLlib	2	2834	April 16, 2021
Entropy Regularization in PG? RLlib	9	880	September 17, 2022
Parameter noise exploration and policy gradient / actor critic RLlib	0	293	September 27, 2022
Adapted GAE formula ==> PPO algorithm used to solve problems modeled as a Semi-Markov Decision Process RLlib	1	323	November 17, 2021
Proper implement of reward scaling in PPO RLlib	0	334	December 17, 2020