Routing policies for a partially observable two-server queueing system.
Wendy EllensPeter KovacsRudesindo Núñez-QueijaHans van den BergPublished in: EAI Endorsed Trans. Ubiquitous Environ. (2016)
Keyphrases
- search algorithm
- partially observable
- markov decision problems
- optimal policy
- markov decision processes
- reward function
- state space
- decision problems
- infinite horizon
- partially observable markov decision processes
- state dependent
- reinforcement learning
- queueing model
- long run
- dynamical systems
- partial observability
- steady state
- belief state
- markov decision process
- partially observable domains
- search space
- scheduling policies
- partial observations
- queue length
- action models
- partially observable environments
- service times
- arrival rate
- linear programming
- single server
- routing protocol
- markov chain
- dynamic programming
- decision processes
- average cost
- sufficient conditions
- orders of magnitude
- supply chain
- special case
- multi agent
- probability distribution
- policy iteration
- random walk
- multiple agents
- planning domains