Methodology of probabilistic analysis of state dynamics of multi­dimensional semi­Markov dynamic systems

Authors

DOI:

https://doi.org/10.15587/1729-4061.2019.184637

Keywords:

dynamic system with many possible states, random transition process, integral dynamic equations, Laplace transforms

Abstract

The problem of probabilistic analysis of a complex dynamic system, which in the process of functioning passes from one state to another at random times, is considered. The methodology for calculating the conditional probabilities of the system getting into a given state at a given time t, provided that at the initial time the system was in any of the possible states is proposed. The initial data for analysis are a set of experimentally obtained values of the duration of the system stay in each of the states before transition to another state. Approximation of the resulting histograms using the Erlang distribution gives a set of distribution densities of the duration of the system stay in possible states before transition to other states. At the same time, the choice of the proper Erlang distribution order provides an adequate description of the semi-Markov processes occurring in the system. The mathematical model that relates the obtained distribution densities to the functions determining the probabilistic dynamics of the system is proposed. The model describes a random process of system transitions from any possible initial state to any other state during a given time interval. Using the model, a system of integral equations for the desired functions describing the probabilistic transition process is obtained. To solve these equations, the Laplace transform is used. As a result of solving the system of integral equations, functions are obtained that specify the probability distribution of the system states at any time t. The same functions also describe the asymptotic probability distribution of states. An illustrative example of solving the problem for the case when the distribution densities of the lengths of the system stay in possible states are described by the second-order Erlang distributions is given. The solution procedure is described in detail for the most natural special case, when the initial state is H0

Author Biographies

Yelyzaveta Meleshko, Central Ukrainian National Technical University Universytetskyi ave., 8, Kropyvnytskyi, Ukraine, 25006

PhD, Associate Professor

Department of Cybersecurity and Software

Lev Raskin, National Technical University "Kharkiv Polytechnic Institute" Kyrpychova str., 2, Kharkiv, Ukraine, 61002

Doctor of Technical Sciences, Professor, Head of Department

Department of Distributed Information Systems and Cloud Technologies

Serhii Semenov, National Technical University "Kharkiv Polytechnic Institute" Kyrpychova str., 2, Kharkiv, Ukraine, 61002

Doctor of Technical Sciences, Professor

Department of Computer Science and Programming

Oksana Sira, National Technical University "Kharkiv Polytechnic Institute" Kyrpychova str., 2, Kharkiv, Ukraine, 61002

Doctor of Technical Sciences, Professor

Department of Distributed Information Systems and Cloud Technologies

References

  1. Berzh, K. (1962). Teoriya grafov i ee prilozheniya. Moscow: IL, 320.
  2. Distel', R. (2002). Teoriya grafov. Novosibirsk: IM, 336.
  3. Tihonov, V. I., Mironov, M. A. (1977). Markovskie protsessy. Moscow: Sovetskoe Radio, 481.
  4. Bulinskiy, A. N., SHiryaev, A. N. (2005). Teoriya sluchaynyh protsessov. Moscow: Fizmatgiz, 364.
  5. Kemeni, Dzh., Snell, Dzh. (1970). Konechnye tsepi Markova. Moscow: Nauka, 198.
  6. Chzhun, K.-L. (1954). Odnorodnye tsepi Markova. Moscow: Mir, 264.
  7. Barucha, R. A. (1969). Elementy teorii Markovskih protsessov. Moscow: Nauka, 320.
  8. Dynkin, E. B. (1963). Markovskie protsessy. Moscow: Fizmatgiz, 482.
  9. Cao, X.-R. (2015). Optimization of Average Rewards of Time Nonhomogeneous Markov Chains. IEEE Transactions on Automatic Control, 60 (7), 1841–1856. doi: https://doi.org/10.1109/tac.2015.2394951
  10. Dimitrakos, T. D., Kyriakidis, E. G. (2008). A semi-Markov decision algorithm for the maintenance of a production system with buffer capacity and continuous repair times. International Journal of Production Economics, 111 (2), 752–762. doi: https://doi.org/10.1016/j.ijpe.2007.03.010
  11. Feinberg, E. A., Yang, F. (2015). Optimal pricing for a GI/M/k/N queue with several customer types and holding costs. Queueing Systems, 82 (1-2), 103–120. doi: https://doi.org/10.1007/s11134-015-9457-7
  12. Li, Q.-L. (2016). Nonlinear Markov processes in big networks. Special Matrices, 4 (1). doi: https://doi.org/10.1515/spma-2016-0019
  13. Li, Q.-L., Lui, J. C. S. (2014). Block-structured supermarket models. Discrete Event Dynamic Systems, 26 (2), 147–182. doi: 10. https://doi.org/10.1007/s10626-014-0199-1
  14. Okamura, H., Miyata, S., Dohi, T. (2015). A Markov Decision Process Approach to Dynamic Power Management in a Cluster System. IEEE Access, 3, 3039–3047. doi: https://doi.org/10.1109/access.2015.2508601
  15. Sanajian, N., Abouee-Mehrizi, H., Balcıog̃lu, B. (2010). Scheduling policies in the M/G/1 make-to-stock queue. Journal of the Operational Research Society, 61 (1), 115–123. doi: https://doi.org/10.1057/jors.2008.139
  16. Krasnov, M. L. (1985). Integral'nye uravneniya. Moscow: Nauka, 476.
  17. Il'in, V. A. (1965). Osnovy matematicheskogo analiza. Moscow: Nauka, 572.
  18. Sveshnikov, A. G., Tihonov, A. N. (1967). Teoriya funktsiy kompleksnoy peremennoy. Moscow: Nauka, 308.

Downloads

Published

2019-11-22

How to Cite

Meleshko, Y., Raskin, L., Semenov, S., & Sira, O. (2019). Methodology of probabilistic analysis of state dynamics of multi­dimensional semi­Markov dynamic systems. Eastern-European Journal of Enterprise Technologies, 6(4 (102), 6–13. https://doi.org/10.15587/1729-4061.2019.184637

Issue

Section

Mathematics and Cybernetics - applied aspects