Point-based Dynamic Programming for DEC-POMDPs

By Daniel Szer, François Charpillet. 2006

93 citations recorded by the source corpus.

How researchers see this paper

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

Citations in the corpus

Multi-agent reinforcement learning as a rehearsal for decentralized planning (2016, critical): “Many exact solvers have been developed for Dec-POMDPs [28,19,26], but these”
Optimal and Approximate Q-value Functions for Decentralized POMDPs (2008, background): “The first of these extensions is due to Szer and Charpillet (2006).”
Heuristic search for identical payoff Bayesian games (2010, methodology): “Since the number of sub-tree policies that are non-dominated tends to be very large, point-based DP (PBDP) methods [10, 8, 3] have been introduced.”
Sample Bounded Distributed Reinforcement Learning for Decentralized POMDPs (2012, methodology): “Because solving Dec-POMDPs is NEXP-complete (Bernstein et al. 2002), exact solu-tion techniques for ﬁnite horizon problems require signiﬁ-cant time and memory resources (Szer and Charpillet 2006; Oliehoek et al. 2010; Spaan, Oliehoek, and Amato 2011).”
Incremental Policy Generation for Finite-Horizon DEC-POMDPs (2009, methodology): “2009) DEC-POMDPs as well as many approximate algorithms such as PBDP (Szer and Charpillet 2006), indefinite-horizon dynamic programming (Amato and Zilberstein 2009) and all algorithms based on MBDP (Seuken and Zilberstein 2007b) (e.”
Improved Memory-Bounded Dynamic Programming for Decentralized POMDPs (2007, methodology): “The Point-Based Dynamic Programming (PBDP) algorithm [Szer and Charpillet, 2006] extends this idea further and computes policies based on a subset of the reachable belief states.”
A Concise Introduction to Decentralized POMDPs (2016, background): “Point-based DP (PBDP) [Szer and Charpillet, 2006] proposes improving pruning of the set Qe,i by considering only a subset of reachable belief points Bi ⊂ (S×Q−i).”
Point-based incremental pruning heuristic for solving finite-horizon DEC-POMDPs (2009, methodology): “Recent attempts have been made to whittle down the set of histories considered [11, 12, 13, 4, 3], but so far the state-of-the-art technique, MBDP [12], still remains constrained by the curse of history.”
Decentralized POMDPs (2012, background): “Point-based DP (PBDP) (Szer and Charpillet, 2006) proposes to improve pruning of the set Q τ=k by considering only a subset B ⊂ P(S× Q−i τ=k) of reachable multi-agent belief points.”