An explicit solution for optimal two-player decentralized control over TCP erasure channels with state feedback
2011pp. 4717–4722
Citations Over TimeTop 22% of 2011 papers
Abstract
We develop an optimal controller synthesis algorithm for decentralized control problems where control actions are transmitted through TCP-like erasure channels. We consider a simple two-player interconnected linear system and Bernoulli distributed erasure channels. We recast the problem to a centralized Partially Observed Markov Decision Process (POMDP) under the fictitious player framework, in which we construct an optimal controller using belief states and value function recursions. Finally, we provide explicit state space formulae for the optimal decentralized controller.
Related Papers
- → Partially Observable Markov Decision Processes(2012)293 cited
- → Information state for Markov decision processes with network delays(2008)9 cited
- → A Method for Speeding Up Value Iteration in Partially Observable Markov Decision Processes(2013)11 cited
- → Markov Decision Processes(2014)6 cited
- → CHQ: a multi-agent reinforcement learning scheme for partially observable markov decision processes(2004)