Optimal Resource Allocation and Policy Formulation in Loosely-Coupled Markov Decision Processes.
Dmitri A. DolgovEdmund H. DurfeePublished in: ICAPS (2004)
Keyphrases
- loosely coupled
- markov decision processes
- optimal policy
- optimal resource allocation
- resource allocation
- policy iteration
- markov decision process
- average reward
- action space
- partially observable
- average cost
- infinite horizon
- finite horizon
- state and action spaces
- reward function
- web services
- decision processes
- state space
- finite state
- reinforcement learning
- distributed systems
- total reward
- dynamic programming
- policy evaluation
- markov decision problems
- service oriented architecture
- reinforcement learning algorithms
- reachability analysis
- decision theoretic planning
- long run
- discounted reward
- decision problems
- transition matrices
- expected reward
- sufficient conditions
- model based reinforcement learning
- control policies
- partially observable markov decision processes
- approximate dynamic programming
- stationary policies
- multistage
- action sets
- initial state
- optimal solution