Planning with Macro|Actions in Decentralized POMDPs

Multi-Agent Planning under Uncertainty with Monte Carlo Q-Value ...

Decentralized partially observable Markov decision processes (Dec-POMDPs) are general multi-agent models for planning under uncertainty, but are intractable ...

Concurrent Reinforcement Learning as a Rehearsal for ...

Decentralized partially-observable Markov decision processes. (Dec-POMDPs) are a powerful tool for modeling multi-agent planning and decision-making under ...

NSF Award Search: Award # 1664923 - CRII: RI: Planning and ...

The PI proposes a theory on macro-actions by using finite-state controllers of Dec-POMDPs. Macro-actions enable the planner to perform multiple ...

Point-Based Backup for Decentralized POMDPs: Complexity and ...

The intractability of optimal POMDP algorithms can be attributed to planning for the complete belief space. DEC-. POMDPs are further disadvantaged as the joint ...

Scalable Planning and Learning for Multiagent POMDPs - DTIC

2008) on large problems by interleaving planning with action selection. The lead ... Decentralized POMDPs. In Wier- ing, M., and van Otterlo, M., eds ...

Lossless Clustering of Histories in Decentralized POMDPs

Decentralized partially observable Markov decision processes ... POMDPs) constitute a generic and expressive framework for multiagent planning under uncertainty.

Knowledge-Based Policies for Qualitative Decentralized POMDPs

We define the planning tasks which we address in two steps; first the model, encoding the dynamics of actions and obser- vations (Definition 2), ...

Scalable and Robust Multi-Agent Reinforcement Learning - Microsoft

• Resulting model: MacDec-POMDP (macro-action Dec-POMDP). Page 15. Macro ... Our decentralized macro-action approach (MA) vs primitive-action version for ...

Monte-Carlo Expectation Maximization for Decentralized POMDPs

In this paper, we focus on infinite-horizon DEC-POMDPs. Our approach is inspired by recent advances in planning by probabilistic inference [Toussaint and ...

Mixed Integer Linear Programming For Exact Finite-Horizon ... - Loria

Partial observability and decentralization make Dec-Pomdps very difficult to solve. Finding an optimal solution to a Dec-. Pomdp is NEXP-hard in the number of ...

DEC‐MDP/POMDP - Markov Decision Processes in Artificial ...

These hypotheses have led to several formalisms. Among them, this chapter reviews the most well-known ones: MMDP, Decentralized MDPs (DEC-MDPs), ...

Multi-agent reinforcement learning as a rehearsal for decentralized ...

Decentralized partially observable Markov decision processes (Dec-POMDPs) are a powerful tool for modeling multi-agent planning and decision ...

Automated Generation of Interaction Graphs for Value-Factored ...

The Decentralized Partially Observable Markov Decision. Process (Dec-POMDP) is a powerful model for multi- agent planning under uncertainty, but its ...

Exact Dynamic Programming for decentralized POMDPs with ...

The reduced policy space contains sequences of actions and observations that are linearly independent. We tested our approach on two benchmark problems, and the ...

Monte-Carlo Planning in Large POMDPs - David Silver

Partially Observable Monte-Carlo Planning (POMCP) consists of a UCT search that selects actions at each time-step; and a particle filter that updates the ...

Exact Dynamic Programming for Decentralized POMDPs with ...

is an efficient POMDP approach that speeds up planning algorithms by ... actions and observations that are linearly independent. We ...

MAGIC: Learning Macro-Actions for Online POMDP Planning

MAGIC: Learning Macro-Actions for Online POMDP Planning. Yiyuan Lee, Panpan Cai and David Hsu. School of Computing, National University of Singapore. Abstract ...

Decentralized Planning under Uncertainty for Teams of ...

As each agent does not know what the other agents observe, it will be hard to predict their actions. Decentralized POMDPs (DEC-. POMDPs) ...

Expectation maximization for average reward decentralized POMDPs

Planning for multiple agents under uncertainty is often based on decentralized partially observable Markov decision processes (Dec-POMDPs), but current ...

Approximate Planning in POMDPs with Macro-Actions - NIPS

Approximate Planning in POMDPs with Macro-Actions. Part of Advances in Neural Information Processing Systems 16 (NIPS 2003) · Bibtex Metadata Paper. Authors.