Compared to policy-based Deep RL algorithms, does this category of algorithms have a lower exploration efficiency ?
I found this information in the following article:
"Optimal energy management strategies for energy internet via deep reinforcement learning approach, Hua et al, 2019" but it didn't cite a source. Is it common knowledge in this expertise ?
Thank you !
Maurice Clerc
Please give a _measure_ of "exploration efficiency" (I mean, not just a vague verbose explanation but a mathematical definition).
Only after that you may be able to study, at least experimentally on some examples and for some algorithms, the evolution of exploration_efficiency vs search_space_size, in order to see if it is a decreasing function or not.
0 votes
0 thanks
More Yasser Hallou's questions See All
Similar questions and discussions