Abstract
Let T and \({\tilde{T}=T-E}\) be arbitrary nonnegative, irreducible, stochastic matrices corresponding to two ergodic Markov chains on n states. A function κ is called a condition number for Markov chains with respect to the (α, β)–norm pair if \({\|\pi-\tilde{\pi}\|_\alpha \leq \kappa(T)\|E\|_\beta}\). Here π and \({\tilde \pi}\) are the stationary distribution vectors of the two chains, respectively. Various condition numbers, particularly with respect to the (1, ∞) and (∞, ∞)-norm pairs have been suggested in the literature. They were ranked according to their size by Cho and Meyer in a paper from 2001. In this paper we first of all show that what we call the generalized ergodicity coefficient \({\tau_p(A^\#)={\rm sup}_{y^{t}e=0} \frac{\|y^tA^\#\|_p}{\|y\|_1}}\), where e is the n-vector of all 1’s and A # is the group generalized inverse of A = I − T, is the smallest condition number of Markov chains with respect to the (p, ∞)-norm pair. We use this result to identify the smallest condition number of Markov chains among the (∞, ∞) and (1, ∞)-norm pairs. These are, respectively, κ 3 and κ 6 in the Cho–Meyer list of 8 condition numbers. Kirkland has studied κ 3(T). He has shown that \({\kappa_3(T)\geq\frac{n-1}{2n}}\) and he has characterized transition matrices for which equality holds. We prove here again that 2κ 3(T) ≤ κ(6) which appears in the Cho–Meyer paper and we characterize the transition matrices T for which \({\kappa_6(T)=\frac{n-1}{n}}\). There is actually only one such matrix: T = (J n − I)/(n − 1), where J n is the n × n matrix of all 1’s.
Similar content being viewed by others
References
Anstreicher K., Rothblum U.: Using Gauss–Jordan elimination to compute the index, generalized nullspace, and Drazin inverse. Lin. Alg. Appl. 85, 221–239 (1987)
Ben-Israel A., Greville T.N.E.: Generalized Inverses: Theory and Applications, 2nd edn. Springer, New York (2003)
Berman A., Plemmons R.J.: Nonnegative Matrices in the Mathematical Sciences. SIAM Publications, Philadelphia (1994)
Campbell S.L., Meyer C.D.: Generalized Inverses of Linear Transformations. Dover Publications, New York (1991)
Cho G.E., Meyer C.D.: Markov chain sensitivity measured by mean first passage times. Lin. Alg. Appl. 316, 21–28 (2000)
Cho G.E., Meyer C.D.: Comparison of perturbation bounds for the stationary distribution of a Markov chain. Lin. Alg. Appl. 335, 137–150 (2001)
Dobrushin, R.L.: Central limit theorem for non-stationary Markov chains, I, II. Theory Probab. Appl. 1, 65–80, 329–383 (1956)
Feller W.: An Introduction to Probability Theory and its Applications, vol. I. Wiley, New York (1960)
Funderlic R., Meyer C.: Sensitivity of the stationary distribution vector for an ergodic Markov chain. Lin. Alg. Appl. 76, 1–17 (1986)
Golub G.H., Meyer C.D.: Using the QR-factorization and group inversion to compute, differentiate, and estimate the sensitivity of stationary probabilities for Markov chains. SIAM J. Alg. Discrete Meth. 7, 273–281 (1986)
Hartfiel D.J., Meyer C.D.: On the structure of stochastic matrices with a subdominant eigenvalue near 1. Lin. Alg. Appl. 272, 193–203 (1998)
Hartwig R.: A method for calculating A d. Math. Japon. 26, 37–43 (1981)
Haviv M., Vander Heyden L.: Perturbation bounds for the stationary probabilities of a finite Markov chain. Adv. Appl. Prob. 16, 804–818 (1984)
Higham N.J., Night P.A.: Componentwise error analysis for stationary iterative methods, in Linear algebra, Markov chains, and queueing models. Inst. Math. Appl. Minnieap. 48, 29–46 (1993)
Ipsen I., Meyer C.: Uniform stability of Markov chains. SIAM J. Matrix Anal. Appl. 15, 1061–1074 (1994)
Kemeny J.G., Snell J.L.: Finite Markov Chains. Van Nostrand, Princeton (1960)
Kirkland S.J.: On a question concerning condition numbers for Markov chains. SIAM J. Matrix Anal. Appl. 23, 1109–1119 (2002)
Kirkland, S.J., Neumann, M.: The case of equality in the Dobrushin–Deutsch–Zenger bound. Lin. Alg. Appl. (2008) (submitted)
Kirkland S., Neumann M., Shader B.: Applications of Pazs inequality to perturbation bounds for Markov chains. Lin. Alg. Appl. 268, 183–196 (1998)
Kirkland S., Neumann M., Xu J.: Transition matrices for well-conditioned Markov chains. Lin. Alg. Appl. 424, 118–131 (2007)
Meyer C.D.: The role of the group generalized inverse in the theory of finite Markov chains. SIAM Rev. 17, 443–464 (1975)
Meyer C.D.: The condition of a finite Markov chain and perturbation bounds for the limiting probabilities. SIAM J. Alg. Discrete Meth. 1, 273–283 (1980)
Meyer C.D.: Sensitivity of the stationary distribution of a Markov chain. SIAM J. Matrix Anal. Appl. 15, 715–728 (1994)
Meyer C.D.: Matrix Analysis and Applied Linear Algebra. SIAM Publications, Philadelphia (2000)
Meyer C.D., Shoaf J.M.: Updating finite Markov chains by using techniques of group matrix inversion. J. Stat. Comput. Simul. 11, 163–181 (1980)
Rothblum U., Tan C.-P.: Upper bounds on the maximum modulus of subdominant eigenvlaues. of nonnegative matrices. Lin. Alg. Appl. 66, 45–86 (1985)
Schweitzer P.: Perturbation theory and finite Markov chains. J. Appl. Probab. 5, 410–413 (1968)
Seneta E.: Non-Negative Matrices and Markov Chains. Springer, New York (1981)
Seneta E.: Perturbation of the stationary distribution measured by ergodicity coefficients. Adv. Appl. Probab. 20, 228–230 (1988)
Seneta E.: Sensitivity analysis, ergodicity coefficients, and rank-one updates for finite Markov chains. In: Stewart, W.J.(eds) Numerical Solution of Markov Chains, Marcel Dekker, New York (1991)
Stewart G.W.: Gaussian elimination, perturbation theory, and Markov chains, in Linear algebra, Markov chains, and queueing models. Inst. Math. Appl. Minneap. 48, 59–69 (1993)
Varga R.S.: Matrix Iterative Analysis, Second Revised and Expanded Edition. Springer, Heidelberg (2000)
Wilkinson J.H.: Note on the practical significance of the Drazin inverse, in Recent applications of generalized inverses. Res. Notes Math. 66, 82–99 (1982)
Author information
Authors and Affiliations
Corresponding author
Additional information
This research was supported in part by NSERC under Grant OGP0138251 and NSA Grant No. 06G–232.
Rights and permissions
About this article
Cite this article
Kirkland, S.J., Neumann, M. & Sze, NS. On optimal condition numbers for Markov chains. Numer. Math. 110, 521–537 (2008). https://doi.org/10.1007/s00211-008-0172-8
Received:
Revised:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s00211-008-0172-8