ABSTRACT
Learning in multiagent systems is generally slow because the agent has to extract its correct policy through not only through its interaction with the environment, but also from its interactions with other learning agents. In this paper, we present an approach that significantly improves the learning speed in multiagent systems by allowing an agent to up-date its estimate of the rewards for all its available actions, not just the action that was taken. Our results show that the rewards on such "actions not taken" are beneficial early in training, particularly when agent teams are leveraged to estimate those rewards.
- A. K. Agogino and K. Tumer. Analyzing and visualizing multiagent rewards in dynamic and stochastic environments. Journal of Autonomous Agents and Multi Agent Systems, 17(2):320--338, 2008. Google ScholarDigital Library
- N. Khani and K. Tumer. Fast Multiagent Learning: Cashing in on Team Knowledge. In Intel. Engr. Systems Though Artificial Neural Nets 18:3--11, 2008.Google Scholar
- P. Stone. Layered Learning in Multi-Agent Systems: A Winning Approach to Robotic Soccer. MIT Press, Cambridge, MA, 2000. Google ScholarDigital Library
- K. Tumer and A. Agogino. Distributed agent-based air traffic flow management. In Proc. of the 6th Intl. Jt. Conf. on Autonomous Agents and Multi-Agent Systems, pp 330--337, Honolulu, May 2007. Best Paper Award. Google ScholarDigital Library
Index Terms
- Learning from actions not taken: a multiagent learning algorithm
Recommendations
Learning in groups of traffic signals
Computer science in general, and artificial intelligence and multiagent systems in particular, are part of an effort to build intelligent transportation systems. An efficient use of the existing infrastructure relates closely to multiagent systems as ...
Learning Sequences of Compatible Actions Among Agents
Action coordination in multiagent systems is a difficult task especially in dynamic environments. If the environment possesses cooperation, least communication, incompatibility and local information constraints, the task becomes even more difficult. Learning ...
Learning about other agents in a dynamic multiagent system
We analyze the problem of learning about other agents in a class of dynamic multiagent systems, where performance of the primary agent depends on behavior of the others. We consider an online version of the problem, where agents must learn models of the ...
Comments