Emerging coordination in infinite team Markov games
نویسندگان
چکیده
In this paper we address the problem of coordination in multi-agent sequential decision problems with infinite statespaces. We adopt a game theoretic formalism to describe the interaction of the multiple decision-makers and propose the novel approximate biased adaptive play algorithm. This algorithm is an extension of biased adaptive play to team Markov games defined over infinite state-spaces. We establish our method to coordinate with probability 1 in the optimal strategy and discuss how this methodology can be combined with approximate learning architectures. We conclude with two simple examples of application of our algorithm.
منابع مشابه
On Asynchronicity of Moves and Coordination∗
This paper shows that asynchronicity of moves can lead to a unique prediction in coordination games, in an infinite-horizon setting, under certain conditions on off-equilibrium payoffs. In two-player games we derive necessary and sufficient conditions for play ultimately being absorbed in the Pareto dominant Nash equilibrium of the stage game, for every Markov perfect equilibrium. For players p...
متن کاملMean Field Asymptotic of Markov Decision Evolutionary Games and Teams
We introduce Markov Decision Evolutionary Games with N players, in which each individual in a large population interacts with other randomly selected players. The states and actions of each player in an interaction together determine the instantaneous payoff for all involved players. They also determine the transition probabilities to move to the next state. Each individual wishes to maximize t...
متن کاملModeling billiards games
Two-player games of billiards, of the sort seen in recent Computer Olympiads held by the International Computer Games Association, are an emerging area with unique challenges for A.I. research. Drawing on the insight gained from our victory in the 2008 ICGA billiards tournament, we define a game-theoretic model of these types of billiards games. The modeling is surprisingly subtle. While sharin...
متن کاملReinforcement Learning to Play an Optimal Nash Equilibrium in Team Markov Games
Multiagent learning is a key problem in AI. In the presence of multiple Nash equilibria, even agents with non-conflicting interests may not be able to learn an optimal coordination policy. The problem is exaccerbated if the agents do not know the game and independently receive noisy payoffs. So, multiagent reinforfcement learning involves two interrelated problems: identifying the game and lear...
متن کاملDynamic Partition of Collaborative Multiagent Based on Coordination Trees
In team Markov games research, it is difficult for an individual agent to calculate the reward of collaborative agents dynamically. We present a coordination tree structure whose nodes are agent subsets or an agent. Two kinds of weights of a tree are defined which describe the cost of an agent collaborating with an agent subset. We can calculate a collaborative agent subset and its minimal cost...
متن کامل