ABSTRACT
We consider the sensitivity of transient solutions of Markov models to perturbations in their generator matrices. The perturbations can either be of a certain structure or can be very general. We consider two different measures of sensitivity and derive upper bounds on them. The derived bounds are sharper than previously reported bounds in the literature. Since the sensitivity analysis of transient solutions is intimately related to the condition of the exponential of the CTMC matrix, we derive an expression for the condition number of the CTMC matrix exponential which leads to some interesting implications. We compare the derived sensitivity bounds both numerically and analytically with those reported in the literature.
- 1.R. Bellman. Introduction to Matriz Analysis, McGraw-Hill, New York, 1969. Google ScholarDigital Library
- 2.R. W. Butler. The Semi-Markov Unreliability Range Evaluator (SURE) Program. NASA Tech. Memo. 86261, 1984.Google Scholar
- 3.Pierre-Jacques Courtois and Pierre Semal. Bounds for Transient Characteristics of Large or Infinite Markov Chains. Numerical Solution of Markov Chains, pp. 413-435, Marcel Dekker Inc., New York, 1991.Google Scholar
- 4.E. de Souza e Silva and H. R. Gail. Calculating availability and performability measures of repairable computer systems using randomization. Journal of the ACM, Vol. 36, No. 1: 171-193, 1989. Google ScholarDigital Library
- 5.J. B. Dugan, K. S. Trivedi, M. K. Smotherman, R. M. Geist. The Hybrid Automated Reliability Predictor. AIAA J. Guidance, Control and Dynamics, Vol. 9, No. 3: 319-331, May-June 1986.Google ScholarCross Ref
- 6.P. M. Frank. Introduction to System Sensitivity. Academic Press, New York, NY, 1978.Google Scholar
- 7.R. E. Funderlic. Sensitivity of the stationary distribution vector for an ergodie Markov chain. Linear Algebra Applications, Vol. 76: 1-17, 1986.Google ScholarCross Ref
- 8.G. H. Golub and C. D. Meyer. Using the QR Faetorization and Group Inversion to Compute, Differentiate, and Estimate the Sensitivity of Stationary Probabilities for Markov Chains. SIAM J. Algebraic and Discrete Methods, Vol. 7, No. 2: 273-281, April 1986. Google ScholarDigital Library
- 9.M. Haviv and L. van der Heyden. Perturbation bounds for the stationary probabilities of a finite Markov chain. Adv. Appl. Probability, Vol. 16: 804- 818, 1984.Google ScholarCross Ref
- 10.P. Heidelberger and A. Goyal. Sensitivity Analysis of Continuous Time Markov Chains Using Uniformization. Computer Performance and Reliability, pp. 93-104, North-Holland, 1987.Google Scholar
- 11.A. 3ensen. Markoff Chains as an Aid in the Study of Markoff Processes. Skand. Aktnarietidskr., Vol. 36: 87-91, 1953.Google Scholar
- 12.T. Kato. Perturbation Theory for Linear Operators, Springer-Verlag, New York, 1966.Google Scholar
- 13.L. D. Lee. Reliability Bounds for Fault-Tolerant Systems with Competing Responses to Component Failures. IEEE Transactions on Reliabilit y, Vol. R- 35: 91-98, 1986.Google ScholarCross Ref
- 14.R. R. Muntz, E. de Souza e Silva and A. Goyal. Bounding Availability of Repairable Computer Systems. IEEE Transactions on Computers, Vol. C-38, No. 12: 1714-1723, Dec. 1989. Google ScholarDigital Library
- 15.Performance Evaluation, Special Issue, February 1992.Google Scholar
- 16.A. Reibman, R. Smith and K. S. Trivedi. Markov and Markov reward model transient analysis: An overview of numerical approaches. European J. of Operations Research, Vol. 40: 257-267, 1989.Google ScholarCross Ref
- 17.J. R. Rice. A Theory of Condition. SlAM J. Numerical Analysis, Vol. 3: 287-310, 1966.Google ScholarCross Ref
- 18.P. J. Schweitzer. Perturbation theory and finite Markov chains. J. AppL Probability, Vol. 5: 401- 413, 1968.Google ScholarCross Ref
- 19.M. Smotherman. Error Analysis in Analytic Reliability Modeling. Microelectronics and Reliability, Vol. 30, No. 1: 141-149, 1990.Google ScholarCross Ref
- 20.M. K. Smotherman, R. M. Geist, K. S. Trivedi. Provably Conservative Approximations to Complex Reliability Models. IEEE Transactions on Computers, Vol. 35, No. 4: 333-338, April 1986. Google ScholarDigital Library
- 21.G. W. Stewart. On the Sensitivity of Nearly Uncoupled Markov Chains. Numerical Solution of Markov Chains, pp. 105-121, Marcel Dekker Inc., New York, 1991.Google Scholar
- 22.K. S. Trivedi. Probability and Statistics with Reliability, Queuing and Computer Science applications. Prentice-Hall, Englewood Cliffs, NJ, 1982. Google ScholarDigital Library
- 23.K. S. Trivedi, J. B. Dugan, R. Geist and M. K. Smotherman. Hybrid Reliability Modeling of Fault- Tolerant Computer Systems. Computers ~ Electrical Engineering, Vol. 11, No. 2/3: 87-108, 1984. Google ScholarDigital Library
- 24.N. M. Van Dijk. Truncation of Markov Chains with Applications to Queueing. Operations Research, Vol. 39, No. 6: 1018-1026, 1991. Google ScholarDigital Library
- 25.N. M. Van Dijk. Perturbation Theory for Markov Reward Processes with Applications to Queueing Systems. Adv. Appi. Probability, Vol. 20: 79-98, 1988.Google ScholarCross Ref
- 26.C. Van Loan. The Sensitvity of the Matrix Exponential. SIAM J. Numerical Analysis, Vol 14, No. 6: 971-981, December 1977.Google ScholarCross Ref
- 27.A. L. White. An Approximation Formula for a Class of Fault-Tolerant Computers. IEEE Transactions on Reliability, Vol. R-35: 99-101, 1986.Google ScholarCross Ref
- 28.W. Whitt. Untold Horrors of the Waiting Room: What the Equilibruim Distribution Will Never Tell About The Queue-Length Process. Management Science, Vol. 5, No. 4: 395-408, April 1983.Google ScholarDigital Library
Index Terms
- On the sensitivity of transient solutions of Markov models
Recommendations
On the sensitivity of transient solutions of Markov models
We consider the sensitivity of transient solutions of Markov models to perturbations in their generator matrices. The perturbations can either be of a certain structure or can be very general. We consider two different measures of sensitivity and derive ...
Multilevel Solutions for Structured Markov Chains
In this paper, a new analysis approach for continuous time Markov chains (CTMCs) with a multidimensional structure is introduced. The presented solution technique employs the multidimensional structure to define aggregated CTMCs that can be analyzed ...
Comments