Abstract
Stochastic processes describe the dynamic behavior of systems modeled by formalisms such as queuing networks, network planning models, semi-Markov processes, and some other. Flowgraph models provide an analytical approach to the problem of stochastic process duration distribution law (SPDDL) finding. The problem is solved in two stages. Initially, the moment generating function (MGF) of the process duration is to be obtained using the graphical evaluation and review technique (GERT) or the flowgraph algebra. This stage is straightforward in contrast to the second one—the analytical transition from the MGF to the process duration distribution law in terms of probability distribution function (cumulative or non-cumulative). The transition is nontrivial and is implemented in this study using MATLAB Symbolic Math Toolbox along with various examples of finding SPDDLs when the processes are represented as ordered activity sets. Also, the capabilities of the statistical flowgraph methodology can be extended over the case when flowgraphs have parallel branches. The results of symbolic calculations are validated via simulation using GPSS World software. This study opens up real possibilities of replacing simulation with symbolic mathematics when searching for duration distribution laws of stochastic processes spawned by flowgraph models.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Similar content being viewed by others
References
Collins, D.H., Warr, R.L., Huzurbazar, A.V.: An introduction to statistical flowgraph models for engineering systems. J. Risk Reliab. 227, 461–470 (2013)
Budiana, M.F., Siddig, M.H.M.A.: Flowgraph models and analysis for Markov jump processes. In: 7th European Business Research Conference, Rome, Italy (2016)
Huzurbazar, A.V.: Flowgraph Models for Multistate Time-to-Event Data. Wiley, Blackwell (2004). https://doi.org/10.1002/0471686565
Rubio, G., García-Mora, B., Santamaría, C., Pontones, J.L.: A flowgraph model for bladder carcinoma. Theor. Biol. Med. Model. 11 (2014). https://doi.org/10.1186/1742-4682-11-S1-S3
Phillips, D.T., Garcia-Diaz, A.: Fundamentals of Network Analysis. Prentice-Hall, Englewood Cliffs, US (1981)
Ren, Y.: The methodology of flowgraph models. Ph.D. thesis, London School of Economics and Political Science, London, England (2011)
Butler, R.W.: Reliabilities for feedback systems and their saddlepoint approximation. Stat. Sci. 15, 279–298 (2000). https://doi.org/10.1214/ss/1009212818
Huzurbazar, A.V.: Flowgraph models for generalized phase type distributions having non-exponential waiting times. Scand. J. Stat. 26, 145–157 (1999). https://doi.org/10.1111/1467-9469.00142
Liang, J., Chen, Y.Q., Guo, B.Z.: A hybrid symbolic-numerical simulation method for some typical boundary control problems. Simulation. 80, 635–643 (2004). https://doi.org/10.1177/0037549704050183
Fomicheva, S.G.: Linear complexity of recurrent sequences. Radiotekhnika 2, 72–77 (1997)
Fatoorehchi, H., Abolghasemi, H.: An integration-free method for inversion of Laplace transforms: a useful tool for process control analysis and design. Chem. Eng. Commun. 203, 822–830 (2016). https://doi.org/10.1080/00986445.2015.1107722
Cheever, E.: The Inverse Laplace Transform by Partial Fraction Expansion. https://lpsa.swarthmore.edu/LaplaceXform/InvLaplace/InvLaplaceXformPFE.html
GPSS World Simulation Environment. https://www.minutemansoftware.com
Sidnev, A.G.: Business process simulation GPSS-models construction (in Russian). In: Senichenkova, Y. (ed.) Vychislitelnye, izmeritelnye i upravliaiushchie sistemy, St. Petersburg, Russia, pp. 41–46 (2006)
Violet UML Editor. https://violet.sourceforge.net
Marsaglia, G., Tsang, W.W., Wang, J.: Evaluating Kolmogorov’s distribution. J. Stat. Softw. 8, 1–4 (2003). https://doi.org/10.18637/jss.v008.i18
Abate, J., Choudhury, G.L., Whitt, W.: An introduction to numerical transform inversion and its application to probability models. In: Grassmann, W.K. (ed.) Computational Probability, pp. 257–323. Springer US, Boston, US (2000). https://doi.org/10.1007/978-1-4757-4828-4_8
Cohen, A.M.: Numerical Methods for Laplace Transform Inversion. Springer, US (2007)
Abate, J., Valkó, P.P.: Multi-precision Laplace transform inversion. Int. J. Numer. Methods Eng. 60, 979–993 (2004). https://doi.org/10.1002/nme.995
Ismail, M.H., Matalgah, M.M.: On the use of Padé approximation for performance evaluation of maximal ratio combining diversity over Weibull fading channels. Eurasip J. Wirel. Commun. Netw. 2006, 1–7 (2005). https://doi.org/10.1155/WCN/2006/58501
Boyd, J.P.: Padé approximant algorithm for solving nonlinear ordinary differential equation boundary value problems on an unbounded domain. Comput. Phys. 11, 299 (1997). https://doi.org/10.1063/1.168606
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Appendix
Appendix
Results for the Uniform Distribution. The Padé approximant orders: a = 6, b = 7. Expectation and variance are presented in Table 4. For comparative charts, see Fig. 11. As a result of the goodness-of-fit test, we failed to reject H0 hypothesis with p = 0.99.
Results for the Triangular Distribution. The Padé approximant orders: a = 6, b = 7. Moment comparison is given in Table 5. For charts of PDFs and CDFs, see Fig. 12. Here, we failed to reject the H0 hypothesis with p = 0.10.
Results for the Erlang-3 Distribution. The Padé approximant orders: a = 6, b = 7. Moment comparison is presented in Table 6. For comparative charts of PDFs and CDFs, see Fig. 13. We failed to reject the H0 hypothesis with p = 0.16.
Results for Concurrent Activities Distributed According to the Uniform Law. Let each of the parallel arcs of activity “34” from Sect. 3.2 be uniformly distributed with the same parameters, and all other distributions of activity durations are also uniform. When simulating in the GPSS World, 5000 iterations were carried out for greater accuracy, and we took a = 12, b = 13 for the Padé approximation.
As a result, we obtained correct values of expectation and variance (see Table 7), and Fig. 14 shows the analytical PDF and CDF compared with the empirical ones constructed from the results of simulation. We failed to reject the H0 hypothesis (p = 0.64).
Results for Concurrent Activities Distributed According to the Exponential Law. The Padé approximant orders: a = 12, b = 13. Moment comparison is presented in Table 8. For comparative charts of PDFs and CDFs, see Figs. 13 and 15. We failed to reject the H0 hypothesis with p = 0.53.
Rights and permissions
Copyright information
© 2021 The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.
About this paper
Cite this paper
Zhemelev, G., Sidnev, A. (2021). Using Symbolic Computing to Find Stochastic Process Duration Distribution Laws. In: Voinov, N., Schreck, T., Khan, S. (eds) Proceedings of International Scientific Conference on Telecommunications, Computing and Control. Smart Innovation, Systems and Technologies, vol 220. Springer, Singapore. https://doi.org/10.1007/978-981-33-6632-9_7
Download citation
DOI: https://doi.org/10.1007/978-981-33-6632-9_7
Published:
Publisher Name: Springer, Singapore
Print ISBN: 978-981-33-6631-2
Online ISBN: 978-981-33-6632-9
eBook Packages: EngineeringEngineering (R0)