Sparse cooperative Q-learning

By Jelle R. Kok, Nikos A. Vlassis. 2004

152 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Molecular Biology

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Molecular Biology

Citations in the corpus

Value Function Transfer for Deep Multi-Agent Reinforcement Learning Based on N-Step Returns (2019, background): “Earlier work focuses on constructing coordination graphs (CGs) to represent the interactive relationship between agents and conducting learning according to the speciﬁed relationship in each state [Guestrin et al. , 2002a; Guestrin et al. , 2002b; Kok and Vlassis, 2004 ] .”
Multi-Agent Common Knowledge Reinforcement Learning (2018, methodology): “Sparse cooperative Q -learning (Kok & Vlassis, 2004) also uses coordination graphs to efﬁ-ciently maximise over joint actions in the Q -learning update rule.”
Factorized Q-learning for large-scale multi-agent systems (2018, background): “gents. Nevertheless, at any particular moment each agent is not supposed to know the other agents’ current action choices, which makes our se−ing much more realistic than a completely centralized one [6]. „e requirement of global knowledge will be further relaxed to the requirement of local neighbor”
A survey and critique of multiagent deep reinforcement learning (2018, background): “There are some hybrid approaches that learn a centralized but factored Q value function [119,174].”
Decentralized MDPs with sparse interactions (2011, methodology): “Coordination graphs have been used for efficient planning and learning in large multiagent MDPs [44].”
Explainable Reinforcement Learning via Reward Decomposition (2019, background): “It turns out that a prior off-policy variant of Qlearning for multi-agent RL [10] can serve our purpose when we interpret agents as reward types.”
Multi-Agent Game Abstraction via Graph Attention Neural Network (2019, methodology): “…techniques such as game abstraction and knowledge transfer to help with speeding up multi-agent re-inforcement learning(Guestrin, Lagoudakis, and Parr 2002; Kok and Vlassis 2004; De Hauwere, Vrancx, and Now ´ e 2010; Melo and Veloso 2011; Hu, Gao, and An 2015; Yu et al. 2015; Liu et al. 2019).”
Collaborative Multiagent Reinforcement Learning by Payoff Propagation (2006, methodology): “In Section 5, we introduce the different variants of our Sparse Cooperative Q-learning method, and give experimental results on several learning problems in Section 6.”
Cooperative Adaptive Cruise Control: A Reinforcement Learning Approach (2011, unknown): “Concerning the space search reduction, Sparse Cooperative Q-Learning [ 10 ] allows agents to coordinate their actions only on predefined set of states.”
Learning multi-agent state space representations (2010, unknown): “Kok & Vlassis [11] explicitly specify these graphs beforehand, while in Kok et al. [10] this approach is extended to also learn the graphs using statistical information about the obtained rewards conditioned on the states and actions of the other agents.”
Multi-agent Reinforcement Learning: An Overview (2010, background): “This direction has been investigated in e.g., [21,38,43,61,62].”
Learning of coordination: exploiting sparse interactions in multiagent systems (2009, background): “In [10,11], the authors further explore the use of coordination graphs, which are now learned from the interactions of the agents.”
Multiagent Learning of Coordination in Loosely Coupled Multiagent Systems (2015, background): “assumptions of agents’ full observability of the joint-state and/or joint-action of other agents [20], [22], [24].”
Multi-objective multi-agent decision making: a utility-based analysis and survey (2019, methodology): “ossibly overlapping) subsets of agents. However, nding suitable joint actions in a MOCoG is also key to nding, e.g., coverage sets for MOMMDPs (as is also the case for single-objective CoGs and MMDPs [51, 52]). In MOCoGs, a lot of research focuses on nding (approximate) Pareto coverage sets (PCSs), ”
Multi-Agent Reinforcement Learning: A Selective Overview of Theories and Algorithms (2019, background): “Several other heuristics (with no theoretical backing) regarding either reward or value function factorization have been proposed to mitigate the scalability issue (Guestrin et al., 2002a,b; Kok and Vlassis, 2004; Sunehag et al., 2018; Rashid et al., 2018).”
Learning in Multi-agent Systems with Sparse Interactions by Knowledge Transfer and Game Abstraction (2015, methodology): “Earlier work makes use of coordination graphs to specify the coordination dependencies between agents in particular states [7, 8, 14].”