On absorption times and Dirichlet eigenvalues

This paper gives a stochastic representation in spectral terms for the absorption time T of a finite Markov chain which is irreducible and reversible outside the absorbing point. This yields quantitative informations on the parameters of a similar representation due to O'Cinneide for general chains admitting real eigenvalues. In the discrete time setting, if the underlying Dirichlet eigenvalues (namely the eigenvalues of the Markov transition operator restricted to the functions vanishing on the absorbing point) are nonnegative, we show that T is distributed as a mixture of sums of independent geometric laws whose parameters are successive Dirichlet eigenvalues (starting from the smallest one). The mixture weights depend on the starting law. This result leads to a probabilistic interpretation of the spectrum, in terms of strong random times and local equilibria through a simple intertwining relation. Next this study is extended to the continuous time framework, where geometric laws have to be replaced by exponential distributions having the (opposite) Dirichlet eigenvalues of the generator as parameters. Returning to the discrete time setting we consider the influence of negative eigenvalues which are given another probabilistic meaning. These results generalize results of Karlin and McGregor and Keilson for birth and death chains.

[1]  Samuel Karlin,et al.  COINCIDENT PROPERTIES OF BIRTH AND DEATH PROCESSES , 1959 .

[2]  Qi-Ming He,et al.  Spectral Polynomial Algorithms for Computing Bi-Diagonal Representations for Phase Type Distributions and Matrix-Exponential Distributions , 2006 .

[3]  Christian Commault,et al.  Phase-type distributions and representations: Some results and open problems for system theory , 2003 .

[4]  Persi Diaconis,et al.  On Times to Quasi-stationarity for Birth and Death Processes , 2009 .

[5]  Colm Art O'Cinneide,et al.  Phase-type distributions and invariant polytopes , 1991, Advances in Applied Probability.

[7]  Marcel F. Neuts,et al.  Matrix-Geometric Solutions in Stochastic Models , 1981 .

[8]  P. Diaconis,et al.  Strong uniform times and finite random walks , 1987 .

[9]  John T. Kent.  The Spectral Decomposition of a Diffusion Hitting Time , 1982 .

[10]  J. A. Fill.  On hitting times and fastest strong stationary times for skip-free chains , 2007 .

[11]  Sivaram K. Narayan,et al.  The nonnegative inverse eigenvalue problem , 2004 .

[12]  J. T. Kent.  Probability, Statistics and Analysis: The appearance of a multivariate exponential distribution in sojourn times for birth-death and diffusion processes , 1983 .

[13]  C. O'Cinneide.  Characterization of phase-type distributions , 1990 .

[14]  James Allen Fill,et al.  The Passage Time Distribution for a Birth-and-Death Chain: Strong Stationary Duality Gives a First Stochastic Proof , 2007, 0707.4042.

[15]  J. Kent.  Eigenvalue expansions for diffusion hitting times , 1980 .

[16]  Persi Diaconis,et al.  Separation cut-offs for birth and death chains , 2006, math/0702411.

[17]  David J. Aldous,et al.  Lower bounds for covering times for reversible Markov chains and random walks on graphs , 1989 .

[18]  James Allen Fill,et al.  On Hitting Times and Fastest Strong Stationary Times for Skip-Free and More General Chains , 2007, 0708.4258.

[19]  Julian Keilson.  Log-concavity and log-convexity in passage time densities of diffusion and birth-death processes , 1971 .

[20]  P. Diaconis,et al.  Strong Stationary Times Via a New Form of Duality , 1990 .

[21]  W. G. Marchal,et al.  Characterizations of generalized hyperexponential distribution functions , 1987 .

[22]  Tosio Kato.  Perturbation theory for linear operators , 1966 .

[23]  Jian Ding,et al.  Total variation cutoff in birth-and-death chains , 2008, 0801.2625.

[24]  C. O'Cinneide.  Phase-type distributions: open problems and a few properties , 1999 .

[25]  J. A. Fill.  Strong stationary duality for continuous-time Markov chains. Part I: Theory , 1992 .

[26]  C. Micchelli,et al.  On functions which preserve the class of Stieltjes matrices , 1979 .

[27]  B. Sengupta.  Matrix geometric solutions in stochastic models , 1987 .

[28]  Qi-Ming He,et al.  PH-Invariant Polytopes and Coxian Representations of Phase Type Distributions , 2006 .

[29]  J. A. Fill.  Time to Stationarity for a Continuous-Time Markov Chain , 1991, Probability in the Engineering and Informational Sciences.