Efficiently Detecting Non-Stationary Opponents: A Bayesian Policy Reuse Approach under Partial Observability

In multi-agent domains, dealing with non-stationary opponents that change behaviors (policies) consistently over time is still a challenging problem, where an agent usually requires the ability to detect the opponent’s policy accurately and adopt the optimal response policy accordingly. Previous wor...

Full beskrivning

Bibliografiska uppgifter

Huvudupphov:	Yu Wang, Ke Fu, Hao Chen, Quan Liu, Jian Huang, Zhongjie Zhang
Materialtyp:	Artigo
Språk:	Inglês
Utgiven:	MDPI AG 2022-07-01
Serie:	Applied Sciences
Ämnen:	Bayesian policy reuse non-stationary opponents variational autoencoders reinforcement learning partial observability
Länkar:	https://www.mdpi.com/2076-3417/12/14/6953
Taggar:	Lägg till en tagg Inga taggar, Lägg till första taggen!