Almost optimal policies for stochastic systemswhich almost satisfy conservation laws.

Kevin D. Glazebrook Renate Garbe

Published in: Ann. Oper. Res. (1999)

Keyphrases

optimal policy
conservation laws
stochastic systems
state dependent
sample path
markov decision processes
lost sales
policy iteration
reinforcement learning

decision problems
finite horizon
average reward
stochastic models
state space
multistage
long run
infinite horizon
dynamic programming

finite state
partial differential equations
initial state
sufficient conditions
long run average cost
demand distributions
least squares
inventory level
markov decision process

partially observable markov decision processes
stochastic model
learning algorithm
monte carlo
confidence intervals
linear programming
denoising
stochastic processes