Zero-Discount Partially Observable Markov Decision Processes For Computational Improvising Agents.
Aengus MartinCraig T. JinAndré van SchaikWilliam L. MartensPublished in: ICMC (2010)
Keyphrases
- partially observable stochastic games
- partially observable markov decision processes
- multi agent
- partial observability
- dynamic programming
- single agent
- finite state
- multi agent systems
- cooperative
- nash equilibrium
- multiple agents
- dynamical systems
- planning under uncertainty
- decision problems
- reinforcement learning
- belief space
- belief state
- multiagent systems
- optimal policy
- continuous state
- dynamic environments
- decision making
- partially observable domains
- markov decision processes
- state space
- decision theoretic
- dec pomdps
- linear programming
- supply chain
- probability distribution
- decentralized control
- fully observable
- stochastic domains
- computational complexity
- learning algorithm
- partially observable markov