An Optimal Best-First Search Algorithm for Solving Infinite Horizon DEC-POMDPs.
Daniel SzerFrançois CharpilletPublished in: ECML (2005)
Keyphrases
- infinite horizon
- dec pomdps
- dynamic programming
- optimal control
- finite horizon
- average cost
- optimal policy
- long run
- optimal plans
- partially observable
- markov decision processes
- theoretical justification
- multi agent
- lead time
- markov decision process
- decision theoretic
- planning under uncertainty
- continuous state
- markov decision problems
- reinforcement learning
- search algorithm for solving
- control strategy
- state space
- policy iteration
- efficient computation
- total cost
- linear programming
- optimal solution
- learning algorithm