Abstract
The distribution of the number of distinct states visited along a Markov chain path is obtained. This random variable is hereby called length of path, to be distinguished from the well known first passage time (the number of steps taken firstly to reach a given state). The length of path is related to a notion of capacity in potential theory for Markov chains.
The use of these results is warranted in a variety of possible applications, wherever probabilistic walks on graphs may be benefically described not only in terms of the number of steps taken, but also by the number of distinct nodes traversed.
In this paper, two different areas of application are considered. One concerns modelling program behaviour in virtual memory systems by a Markov chain model. In this context, the paging rate of the Least Recently Used paging algorithm is obtained. The other concerns the memory bandwidth in interleaved memory systems with saturated demand, where the stream of memory requests is Markovian. An approach to computing the mean memory bandwidth is proposed.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Preview
Unable to display preview. Download preview PDF.
References
Aho, A. V., P. J. Denning, J. D. Ullman, “Principles of Optimal Page Replacement,” J. Assoc. Comput. Mach., 18, (1971).
Burnett, G. J., E. G. Coffman, “A Combinatorial Problem Related to Interleaved Memory Systems,” J. Assoc. Comput. Mach., 20, (1973), 34–95.
Courtois, P. J., Decomposability, Academic Press, 1977.
Courtois, P. J., H. Vantilborgh, “A Decomposable Model of Program Paging Behaviour,” Acta Inform., 6, (1976), 251–275.
Denning, P. J., S. C. Schwartz, “Properties of the Working Set Model,” Comm. ACM, 15, (1972), 191–198.
Feller, W., An Introduction to Probability Theory and its Application, I, 3rd ed., John Wiley, 1968.
Franklin, M. A., R. K. Gupta, “Computation of Page Fault Probabilities from Program Transition Diagrams,” Comm. ACM, 17, (1974), 186–191.
Freiberger, W. F., U. Grenander, P. D. Sampson, “Patterns in Program Behaviour,” IBM J. Res. Develop., 19, (1975), 230–243.
Gelenbe, E., “A Unified Approach to the Evaluation of a Class of Replacement Algorithms,” IEEE Trans. Comput., C-226, (1973), 611–618.
Glowacki, C., “A Closed form Expression of the Page Fault Rate for the LRU Algorithm in a Markovian Reference Model of Program Behaviour,” in Proc. Int. Comp. Symp., E. Morlet, D. Ribbens (Eds.), North Holland, 1977, 315-318.
Hofri, M., P. Tzelnic, “On the Working Set size for the Markov Chain Model of Program Behaviour,” in Performance of Comp. Syst., M. Arato et al. (Eds.), North Holland, 1979, 393-405.
Kemeny, J. G., J. L. Snell, Finite Markov Chains, Van Nostrand, 1960.
Syski, R., “Potential Theory for Markov Chains,” in Probab. Methods in Appl. Math., III, A. T. Barucha-Reid (Ed.), Academic Press, 1973.
Syski, R., personal communication.
Török, T. L., “Interleaved Memory Systems with Markovian Requests,” in Performance of Comp. Syst., M. Arato et al. (Eds.), North Holland, 1979.
Tzelnic, P., “Stochastic Models of Program Behaviour in Paged Virtual Memory Systems,” D. Sc. Thesis (Hebr.), Department of Computer Science, Israel Institute of Technology, 1979.
Endnotes
Strictly speaking, the first hit time must be defined in equation (2.4) for t ≥ 1, and so that X(1) ε Hc. Equation (2.4), with X(1) ε H, is in fact the definition of the first return time to H; but the distinction between these two variables is of no import in the sequel.
It is for this reason that H-open and H-closed paths are not the same as, respectively, first hit and first return paths (as mentioned in the note on page 3.)
Author information
Authors and Affiliations
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 1982 Springer Science+Business Media New York
About this chapter
Cite this chapter
Tzelnic, P. (1982). The Length of Path for Finite Markov Chains and its Application to Modelling Program Behaviour and Interleaved Memory Systems. In: Disney, R.L., Ott, T.J. (eds) Applied Probability— Computer Science: The Interface. Progress in Computer Science, vol 3. Birkhäuser, Boston, MA. https://doi.org/10.1007/978-1-4612-5798-1_17
Download citation
DOI: https://doi.org/10.1007/978-1-4612-5798-1_17
Publisher Name: Birkhäuser, Boston, MA
Print ISBN: 978-0-8176-3093-5
Online ISBN: 978-1-4612-5798-1
eBook Packages: Springer Book Archive