Abstract
Stochastic max-plus linear systems, i.e., perturbed systems that are linear in the max-plus algebra, belong to a special class of discrete-event systems that consists of systems with synchronization but no choice. In this paper, we study the identification problem for such systems, considering two different approaches. One approach is based on exact computation of the expected values and consists in recasting the identification problem as an optimization problem that can be solved using gradient-based algorithms. However, due to the structure of stochastic max-plus linear systems, this method results in a complex optimization problem. The alternative approach discussed in this paper, is an approximation method based on the higher-order moments of a random variable. This approach decreases the required computation time significantly while still guaranteeing a performance that is comparable to the one of the exact solution.
Similar content being viewed by others
Notes
Intuitively, this can be characterized as follows. Note that Eqs. 3 and 4 imply that each component of x(k + 1) can be written as a max expression of terms in which the unknown parameters θ and λ appear. An input signal is then said to be sufficiently rich if it is such that each of these terms is the maximal one sufficiently often (this is also related to the idea of persistent excitation in conventional system identification (Ljung 1999)).
If there are two identical affine arguments in the max expression in Eq. 10, then the corresponding sets Ω ij coincide. So in general the Ω ij sets either coincide or they only overlap at the boundaries (see also Remark 1). For the sake of simplicity of the exposition, we assume here that any identical arguments in the max expression in Eq. 10 have already been eliminated.
For the case that f(e) is a continuous function, \(\hat{\eta}_i(k+1,\hat{\theta},\hat{\lambda})\) is continuously differentiable and since it is a convex function, the subgradients are unique and they are equal to the gradients.
In the rest of this section, the index k will be dropped (except for η i and \(\hat \eta_i\)) for the sake of simplicity of notation.
These times were obtained running Matlab 7.5.0 (R2007b) on a 2.33 GHz Intel Core Duo E655 processor.
The relative error is defined as \(\frac{|x_0 - x|}{|x|}\) where x is the true value and x 0 is the estimated value.
Note that here due to the 3σ-rule, we choose λ one third of the one in Example 7.1.
References
Abramowitz MA, Stegun I (1964) Handbook of mathematical functions. National Bureau of Standards, US Government Printing Office, Washington DC
Akian M (2007) Representation of stationary solutions of Hamilton-Jacobi-Bellman equations: a max-plus point of view. In: Proceedings of the international workshop “Idempotent and Tropical Mathematics and Problems of Mathematical Physics”. Moscow, Russia
Baccelli F, Cohen G, Olsder GJ, Quadrat JP (1992) Synchronization and linearity. Wiley, New York
Başar T, Bernhard P (1995) H∞-optimal control and related minimax design problems, 2nd edn. Birkhauser, Boston, Massachusetts
Bemporad A, Borrelli F, Morari M (2003) Min-max control of constrained uncertain discrete-time linear systems. IEEE Trans Automat Contr 48(9):1600–1606
Boimond JL, Hardouin L, Chiron P (1995) A modeling method of SISO discrete-event systems in max-algebra. In: Proceedings of the 3rd European control conference. Rome, Italy, pp 2023–2026
Boyd S, Vandenberghe L (2004) Convex optimization. Cambridge University Press, London, UK
Büeler B, Enge A, Fukuda K (2000) Exact volume computation for convex polytopes: a practical study. In: Kalai G, Ziegler G (eds) Polytopes–combinatorics and computation. Birkäuser Verlag, Basel, Switzerland, pp 131–154
Cassandras CG, Lafortune S (1999) Introduction to discrete event systems. Kluwer Academic Publishers, Boston, Massachusetts
Cuninghame-Green RA (1979) Minimax algebra. Springer-Verlag, Berlin, Germany
Davis PJ, Rabinowitz P (1984) Methods of numerical integration, 2nd edn. Academic Press, New York
De Schutter B, van den Boom T, Verdult V (2002) State space identification of max-plus-linear discrete event systems from input-output data. In: Proceedings of the 41st IEEE conference on decision and control. Las Vegas, Nevada, pp 4024–4029
Dekking FM, Kraaikamp C, Lopuhaä HP, Meester LE (2005) A modern introduction to probability and statistics. Springer, London, UK
Farahani SS, van den Boom T, van der Weide H, De Schutter B (2010) An approximation approach for model predictive control of stochastic max-plus linear systems. In: Proceedings of the 10th international workshop on discrete event systems (WODES’10). Berlin, Germany, pp 386–391
Gallot F, Boimond JL, Hardouin L (1997) Identification of simple elements in max-algebra: application to SISO discrete event systems modelisation. In: Proceedings of the European control conference (ECC’97). Brussels, Belgium, paper 488
Golub GH, Van Loan CF (1996) Matrix computations, 3rd edn. The John Hopkins University Press, Baltimore, Maryland
Goodwin GC, Payne RL (1977) Dynamic system identification: experiment design and data analysis. Academic Press, New York
Graham R, Knuth D, Patashnik O (1994) Concrete mathematics: a foundation for computer science, 2nd edn. Addison-Wesley, Boston, Massachusetts
Heidergott B, Olsder GJ, van der Woude J (2006) Max plus at work. Princeton University Press, Princeton, New Jersey
Ho YC (ed) (1992) Discrete event dynamic systems: analyzing complexity and performance in the modern world. IEEE Press, Piscataway, New Jersey
Ivelić S, Pečarić JE (2011) Remarks on “on a converse of Jensen’s discrete inequality” of S. Simić. Hindawi Publishing Corporation, Journal of Inequalities and Applications Article ID 309565
Kalos MH, Whitlock PA (2008) Monte Carlo methods. Wiley-VCH, Weinheim, Germany
Lasserre JB (1998) Integration on a convex polytope. Proc Am Math Soc 126(8):2433–2441
Ljung L (1999) System identification: theory for the user, 2nd edn. Prentice-Hall, Upper Saddle River, New Jersey
Mairesse J (1994) Stochastic linear systems in the (max,+) algebra. In: 11th international conference on analysis and optimization of systems discrete event systems. Lecture notes in control and information sciences, vol 199. Springer-Verlag, London, UK, pp 257–265
Mattheiss TH, Rubin DS (1980) A survey and comparison of methods for finding all vertices of convex polyhedral sets. Math Oper Res 5(2):167–185
McEneaney WM (2004) Max-plus eigenvector methods for nonlinear H-infinity problems: error analysis. SIAM J Control Optim 43:379–412
Menguy E, Boimond JL, Hardouin L, Ferrier JL (2000) A first step towards adaptive control for linear systems in max algebra. Discret Event Dyn Syst: Theory Appl 10(4):347–367
Olsder GJ, Resing JAC, de Vries RE, Keane MS, Hooghiemstra G (1990) Discrete event systems with stochastic processing times. IEEE Trans Automat Contr 35(3):299–302
Papoulis A (1991) Probability, random variables, and stochastic processes. McGraw-Hill, Singapore
Pardalos PM, Resende MGC (eds) (2002) Handbook of applied optimization. Oxford University Press, Oxford, UK
Peterson JL (1981) Petri net theory and the modeling of systems. Prentice-Hall, Englewood Cliffs, New Jersey
Pečarić JE, Beesack PR (1987) On knopp’s inequality for convex functions. Can Math Bull 30(3):267–272
Resing JAC, de Vries RE, Hooghiemstra G, Keane MS, Olsder GJ (1990) Asymptotic behavior of random discrete event systems. Stoch Process Their Appl 36(2):195–216
Rudin W (1987) Real and complex analysis, 3rd edn. McGraw-Hill Series in Higher Mathematics, McGraw-Hill, New York
Schullerus G, Krebs V (2001a) Diagnosis of batch processes based on parameter estimation of discrete event models. In: Proceedings of the European control conference 2001 (ECC’01). Porto, Portugal, pp 1612–1617
Schullerus G, Krebs V (2001b) Input signal design for discrete event model based batch process diagnosis. In: Proceedings of the 4th IFAC workshop on on-line fault detection and supervision in the chemical process industries. Seoul, Korea
Schullerus G, Krebs V, De Schutter B, van den Boom T (2003) Optimal input signal design for identification of max-plus-linear systems. In: Proceedings of the 2003 European control conference (ECC’03). Cambridge, UK, paper 026
Simić S (2009a) On a converse of Jensen’s discrete inequality. J Inequal Appl 2009(1):153080
Simić S (2009b) On an upper bound for Jensen’s inequality. Journal of Inequalities in Pure and Applied Mathematics 10(2), Article 60
Somasundaram KK, Baras JS (2011) Solving multi-metric network problems an interplay between idempotent semiring rules. Linear Algebra Appl 435(7):1494–1512
van den Boom T, De Schutter B (2004) Model predictive control for perturbed max-plus-linear systems: a stochastic approach. Int J Control 77(3):302–309
van den Boom TJJ, De Schutter B, Verdult V (2003) Identification of stochastic max-plus-linear systems. In: Proceedings of the 2003 European control conference (ECC’03). Cambridge, UK, paper 104
Willink R (2005) Normal moments and Hermite polynomials. Stat Probab Lett 73(3):271–275
Acknowledgements
The authors would like to thank Dr. Ioan Landau for his useful comments and suggestions and Dr. Hans van der Weide for his help in the derivation of the approximation method presented in Farahani et al. (2010). This research is partially funded by the Dutch Technology Foundation STW project “Model-predictive railway traffic management” (11025), and by the European Union Seventh Framework Programme [FP7/2007-2013] under grant agreement no. 257462 HYCON2 Network of Excellence.
Author information
Authors and Affiliations
Corresponding author
Rights and permissions
About this article
Cite this article
Farahani, S.S., van den Boom, T. & De Schutter, B. Exact and approximate approaches to the identification of stochastic max-plus-linear systems. Discrete Event Dyn Syst 24, 447–471 (2014). https://doi.org/10.1007/s10626-013-0164-4
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s10626-013-0164-4