ConceptIndependent Q-Learning / IQL

Independent Q-Learning

A simple multi-agent reinforcement learning method in which each agent independently treats the other agents as part of the environment and runs single-agent Q-learning, ignoring the non-stationarity caused by interaction.

Timeline (last 90 days)

Oct 18

Independent Q-Learning, though simple, ignores non-stationarity and easily leads to unstable learning.

Unverified50%

All Facts (1)

Unverified

Independent Q-Learning, though simple, ignores non-stationarity and easily leads to unstable learning.

50%

Source Articles

The Challenge of Interaction-Mediated Nonstationarity in Multi-Agent Reinforcement LearningOct 18