Abstract
The standard coupling from the past (CFTP) algorithm is an interesting tool to sample from exact Markov chain steady-state probability. The CFTP detects, with probability one, the end of the transient phase (called burn-in period) of the chain and consequently the beginning of its stationary phase. For large and/or stiff Markov chains, the burn-in period is expensive in time consumption. In this work, we propose a kind of dual form for CFTP called D-CFTP that, in many situations, reduces the Monte Carlo simulation time and does not need to store the history of the used random numbers from one iteration to another. A performance comparison of CFTP and D-CFTP will be discussed, and some numerical Monte Carlo simulations are carried out to show the smooth running of the proposed D-CFTP.
Funding statement: Supported by Ibn-al-Banna Laboratory of Math. and Appl. (LIBMA), Cadi Ayyad University.
References
[1] M. Y. Bounnite and A. Nasroallah, Widening and clustering techniques allowing the use of monotone CFTP algorithm, Monte Carlo Methods Appl. 21 (2015), no. 4, 301–312. 10.1515/mcma-2015-0111Search in Google Scholar
[2] Y. Cai, A non-monotone CFTP perfect simulation method, Statist. Sinica 15 (2005), no. 4, 927–943. Search in Google Scholar
[3] G. Casella, K. L. Mengersen, C. P. Robert and D. M. Titterington, Perfect samplers for mixtures of distributions, J. R. Stat. Soc. Ser. B Stat. Methodol. 64 (2002), no. 4, 777–790. 10.1111/1467-9868.00360Search in Google Scholar
[4] S. Connor, Coupling: Cutoffs, CFTP and tameness, Ph.D. thesis, University of Warwick, 2007. Search in Google Scholar
[5] S. B. Connor and W. S. Kendall, Perfect simulation for a class of positive recurrent Markov chains, Ann. Appl. Probab. 17 (2007), no. 3, 781–808. 10.1214/105051607000000032Search in Google Scholar
[6] S. N. Ethier and T. G. Kurtz, Markov Processes. Characterization and Convergence, Wiley Ser. Probab. Stat., John Wiley & Sons, New York, 1986. 10.1002/9780470316658Search in Google Scholar
[7] H. Fakhouri, Estimation du régime stationnaire de modèles markoviens complexes par l’algorithme de couplage depuis le passé, Thèse de doctorat, Cadi Ayyad University, 2010. Search in Google Scholar
[8] H. Fakhouri and A. Nasroallah, On the simulation of Markov chain steady-state distribution using CFTP algorithm, Monte Carlo Methods Appl. 15 (2009), no. 2, 91–105. 10.1515/MCMA.2009.005Search in Google Scholar
[9] J. A. Fill, An interruptible algorithm for perfect sampling via Markov chains, Ann. Appl. Probab. 8 (1998), no. 1, 131–162. 10.1214/aoap/1027961037Search in Google Scholar
[10] J. A. Fill, M. Machida, D. J. Murdoch and J. S. Rosenthal, Extension of Fill’s perfect rejection sampling algorithm to general chains, Random Structures Algorithms 17 (2000), no. 3–4, 290–316. 10.1002/1098-2418(200010/12)17:3/4<290::AID-RSA6>3.0.CO;2-QSearch in Google Scholar
[11] S. G. Foss and R. L. Tweedie, Perfect simulation and backward coupling, Commun. Stat. Stochastic Models 14 (1998), no. 1–2, 187–203. 10.1080/15326349808807466Search in Google Scholar
[12] W. S. Kendall and J. Mø ller, Perfect simulation using dominating processes on ordered spaces, with application to locally stable point processes, Adv. in Appl. Probab. 32 (2000), no. 3, 844–865. 10.1239/aap/1013540247Search in Google Scholar
[13] A. Mira, J. Mø ller and G. O. Roberts, Perfect slice samplers, J. R. Stat. Soc. Ser. B Stat. Methodol. 63 (2001), no. 3, 593–606. 10.1111/1467-9868.00301Search in Google Scholar
[14] J. Mø ller, Perfect simulation of conditionally specified models, J. R. Stat. Soc. Ser. B Stat. Methodol. 61 (1999), no. 1, 251–264. 10.1111/1467-9868.00175Search in Google Scholar
[15] D. J. Murdoch and P. J. Green, Exact sampling from a continuous state space, Scand. J. Statist. 25 (1998), no. 3, 483–502. 10.1111/1467-9469.00116Search in Google Scholar
[16] D. J. Murdoch and J. S. Rosenthal, Efficient use of exact samples, Stat. Comput. 10 (1999), 237–243. 10.1023/A:1008991527785Search in Google Scholar
[17] A. Philippe and C. Robert, Perfect simulation of positive Gaussian distributions, Stat. Comput. 13 (2003), no. 2, 179–186. 10.1023/A:1023264710933Search in Google Scholar
[18] J. G. Propp and D. B. Wilson, Exact sampling with coupled Markov chains and applications to statistical mechanics, Random Structures Algorithms 9 (1996), no. 1–2, 223–252. 10.1002/(SICI)1098-2418(199608/09)9:1/2<223::AID-RSA14>3.0.CO;2-OSearch in Google Scholar
[19] J. G. Propp and D. B. Wilson, How to get a perfectly random sample from a generic Markov chain and generate a random spanning tree of a directed graph, J. Algorithms 27 (1998), no. 2, 170–217. 10.1006/jagm.1997.0917Search in Google Scholar
[20] D. B. Wilson, How to couple from the past using a read-once source of randomness, Random Structures Algorithms 16 (2000), no. 1, 85–113. 10.1002/(SICI)1098-2418(200001)16:1<85::AID-RSA6>3.0.CO;2-HSearch in Google Scholar
[21] D. B. Wilson, Layered multishift coupling for use in perfect sampling algorithms (with a primer on CFTP), Monte Carlo Methods (Toronto 1998), Fields Inst. Commun. 26, American Mathematical Society, Providence (2000), 143–179. 10.1090/fic/026/11Search in Google Scholar
© 2019 Walter de Gruyter GmbH, Berlin/Boston
