Speed of stability for birth-death processes

Abstract

This paper is a continuation of the study on the stability speed for Markov processes. It extends the previous study of the ergodic convergence speed to the non-ergodic one, in which the processes are even allowed to be explosive or to have general killings. At the beginning stage, this paper is concentrated on the birth-death processes. According to the classification of the boundaries, there are four cases plus one more having general killings. In each case, some dual variational formulas for the convergence rate are presented, from which, the criterion for the positivity of the rate and an approximating procedure of estimating the rate are deduced. As the first step of the approximation, the ratio of the resulting bounds is usually no more than 2. The criteria as well as basic estimates for more general types of stability are also presented. Even though the paper contributes mainly to the non-ergodic case, there are some improvements in the ergodic one. To illustrate the power of the results, a large number of examples are included.

Keywords

Birth-death process / speed of stability / first eigenvalue / variational formula / criterion and basic estimates / approximating procedure / duality / killing

References

[1] Chen A., Pollett P., Zhang H., Cairns B. Uniqueness criteria for continuous-time Markov chains with general transition structure. Adv Appl Prob, 2005, 37(4): 1056-1074.

[2] Chen M. F. Exponential L²-convergence and L²-spectral gap for Markov processes. Acta Math Sin (New Ser), 1991, 7(1): 19-37.

[3] Chen M. F. Estimation of spectral gap for Markov chains. Acta Math Sin (New Ser), 1996, 12(4): 337-360.

[4] Chen M. F. Analytic proof of dual variational formula for the first eigenvalue in dimension one. Sci in China, A, 1999, 42(8): 805-815.

[5] Chen M. F. The principal eigenvalue for jump processes. Acta Math Sin (Eng Ser), 2000, 16(3): 361-368.

[6] Chen M. F. Explicit bounds of the first eigenvalue. Sci Chin, Ser A, 2000, 43(10): 1051-1059.

[7] Chen M. F. Variational formulas and approximation theorems for the first eigenvalue in dimension one. Sci Chin, Ser A, 2001, 44(4): 409-418.

[8] Chen M F. Ergodic Convergence Rates of Markov Processes-Eigenvalues, Inequalities and Ergodic Theory. 2001. [Collection of papers, 1993-] http://math.bnu.edu.cn/~chenmf/main_eng.htm

[9] Chen M. F. Variational formulas of Poincaré-type inequalities for birth-death processes. Acta Math Sin (Eng Ser), 2003, 19(4): 625-644.

[10] Chen M. F. From Markov Chains to Non-equilibrium Particle Systems, 2004 2nd ed. Singapore: World Scientific.

[11] Chen M. F. Capacitary criteria for Poincaré-type inequalities. Potential Theory, 2005, 23(4): 303-322.

[12] Chen M. F. Eigenvalues, Inequalities, and Ergodic Theory, 2005, London: Springer.

[13] Chen M. F., Wang F. Y. Estimation of spectral gap for elliptic operators. Trans Amer Math Soc, 1997, 349(3): 1239-1267.

[14] Chen M. F., Wang F. Y. Cheeger’s inequalities for general symmetric forms and existence criteria for spectral gap. (Abstract) Chin Sci Bull, 1998, 43(18): 1516-1519.

[15] Chen M. F., Zhang Y. H., Zhao X. L. Dual variational formulas for the first Dirichlet eigenvalue on half-line. Sci China, Ser A, 2003, 46(6): 847-861.

[16] Cox J. T., Rösler U. A duality relation for entrance and exit laws for Markov processes. Stoch Proc Appl, 1983, 16: 141-156.

[17] Dobrushin R. L. On conditions of regularity of stationary Markov processes with a denumerable number of possible states. Uspehi Matem Nauk (NS), 1952, 7(6): 185-191.

[18] Fukushima M., Uemura T. Capacitary bounds of measures and ultracontractivity of time changed processes. J Math Pure et Appliquees, 2003, 82(5): 553-572.

[19] Hou Z. T., Liu Z. M., Zhang H. J., Li J. P., Zhou J. Z., Yuan C. G. Birth-death Processes, 2000, Changsha: Hunan Sci and Tech Press.

[20] Hou Z. T., Zhou J. Z., Zhang H. J., Liu Z. M., Xiao G. N., Chen A. Y., Fei Z. L. The Q-matrix Problem for Markov Chains, 1994, Changsha: Hunan Sci and Tech Press.

[21] Jin H. Y. Estimating the constant in Lp-Poincaré inequality. Master Thesis, 2006, Beijing: Beijing Normal Univ.

[22] Karlin S., McGregor J. L. The differential equations of birth-and-death processes, and the Stieltjes moment problem. Trans Amer Math Soc, 1957, 85: 589-646.

[23] Karlin S., McGregor J. The classification of birth and death processes. Trans Amer Math Soc, 1957, 86(2): 366-400.

[24] Kijima M. Markov Processes for Stochastic Modeling, 1997, London: Chapman & Hall.

[25] Mao Y. H. Nash inequalities for Markov processes in dimension one. Acta Math Sin (Eng Ser), 2002, 18(1): 147-156.

[26] Mao Y. H., Xia L. Y. Spectral gap for jump processes by decomposition method. Front Math China, 2009, 4(2): 335-347.

[27] Maz’ya V. G. Sobolev Spaces, 1985, Berlin: Springer.

[28] Miclo L. An example of application of discrete Hardy’s inequalities. Markov Processes Relat Fields, 1999, 5: 319-330.

[29] Muckenhoupt B. Hardy’s inequality with weights. Studia Math, 1972, XLIV: 31-38.

[30] Opic B., Kufner A. Hardy-type Inequalities, 1990, New York: Longman.

[31] Shao J. H., Mao Y. H. Estimation of the Dirichlet eigenvalue of birth-death process on trees. Acta Math Sin (Chinese Ser), 2007, 50(3): 507-516.

[32] Shiozawa Y., Takeda M. Variational formula for Dirichlet forms and estimates of principal eigenvalues for symmetric α-stable processes. Potential Analysis, 2005, 23: 135-151.

[33] Sirl D., Zhang H., Pollett P. Computable bounds for the decay parameter of a birth-death process. J Appl Prob, 2007, 44(2): 476-491.

[34] van Doorn E. A. Stochastic Monotonicity and Queuing Applications of Birth-Death Processes. Lecture Notes in Statistics, 1981, Berlin: Springer-Verlag.

[35] van Doorn E. A. Conditions for exponential ergodicity and bounds for the decay parameter of a birth-death process. Adv Appl Prob, 1985, 17: 514-530.

[36] van Doorn E. A. Representations and bounds for zeros of orthogonal polynomials and eigenvalues of sign-symmetric tri-diagonal matrices. J Approx Th, 1987, 51: 254-266.

[37] van Doorn E. A. Representations for the rate of convergence of birth-death processes. Theory Probab Math Statist, 2002, 65: 37-43.

[38] Wang J. First Dirichlet eigenvalue of transient birth-death processes. 2008, preprint

[39] Wang J. Poincaré-type inequalities for transient birth-death processes. 2008, preprint

[40] Wang J. Functional inequalities for transient birth-death processes and their applications. 2008, preprint

[41] Wang Z. K. The ergodicity and zero-one law for birth and death processes. Acta Sci Nankai Univ, 1964, 5(5): 93-102.

[42] Wang Z. K., Yang X. Q. Birth and Death Processes and Markov Chains, 1992, Berlin: Springer.

[43] Zeifman A. I. Some estimates of the rate of convergence for birth and death processes. J Appl Prob, 1991, 28: 268-277.

[44] Zhang X. The Estimation for the First Eigenvalue of Schrödinger Operators and a Class of Geometric Inequalities. Ph D Thesis, 2007, Beijing: Beijing Normal University.