all AI news
Satisficing Paths and Independent Multi-Agent Reinforcement Learning in Stochastic Games. (arXiv:2110.04638v2 [cs.GT] UPDATED)
cs.LG updates on arXiv.org arxiv.org
In multi-agent reinforcement learning (MARL), independent learners are those
that do not access the action selections of other agents in the system. Due to
the decentralization of information, it is generally difficult to design
independent learners that drive the system to equilibrium. This paper
investigates the feasibility of using \emph{satisficing} dynamics to guide
independent learners to approximate equilibrium policies in non-episodic,
discounted stochastic games. Satisficing refers to halting search in an
optimization problem upon finding a satisfactory but possibly suboptimal …
arxiv games independent learning reinforcement learning stochastic