A kind of dual form for coupling from the past algorithm, to sample from Markov chain steady-state probability.
Abdelaziz NasroallahMohamed Yasser BounnitePublished in: Monte Carlo Methods Appl. (2019)
Keyphrases
- steady state
- markov chain
- monte carlo
- markov model
- monte carlo simulation
- learning algorithm
- transition probabilities
- product form
- monte carlo method
- queue size
- dynamic programming
- markov processes
- markov process
- finite state
- expectation maximization
- stationary distribution
- random walk
- transition matrix
- queue length
- neural network
- queueing networks
- stochastic process
- algo rithm
- state dependent
- probabilistic model
- parameter estimation
- queueing model
- service times
- arrival rate
- bayesian networks
- confidence intervals
- explicit expressions