Reduction of acyclic phase-type representations

Acyclic phase-type distributions are phase-type distributions with triangular matrix representations. They constitute a versatile modelling tool, since they (1) can serve as approximations to any continuous probability distribution, and (2) exhibit special properties and characteristics that usually make their analysis easier. The size of the matrix representations has a strong effect on the computational efforts needed in analyzing these distributions. These representations, however, are not unique, and two representations of the same distribution can differ drastically in size. This thesis proposes an effective algorithm to reduce the size of the matrix representations without altering their associated distributions. The algorithm produces significantly better reductions than existing methods. Furthermore, the algorithm consists in only standard numerical computations, and therefore is straightforward to implement. We identify three operations on acyclic phase-type representations that arise often in stochastic models. Around these operations we develop a simple stochastic process calculus, which provides a framework for stochastic modelling and analysis. We prove that the representations produced by the three operations are "almost surely" minimal, and the reduction algorithm can be used to obtain these almost surely minimal representations. The applicability of these contributions is exhibited on a variety of case studies. Azyklische Phasentypverteilungen sind Phasentypverteilungen, deren Matrixdarstellung eine Dreiecksmatrix ist. Sie stellen ein vielseitiges Modellierungswerkzeug dar, da sie einerseits als Approximationen jeder beliebigen kontinuierlichen Wahrscheinlichkeitsverteilung dienen konnen, und andererseits spezielle Eigenschaften und Charakteristiken aufweisen, die ihre Analyse vereinfachen. Die Grose der Matrixdarstellung hat dabei einen starken Einfluss auf den Berechnungsaufwand, der zur Analyse solcher Verteilungen notig ist. Die Matrixdarstellung ist jedoch nicht eindeutig, und zwei verschiedene Darstellungen ein und derselben Verteilung konnen sich drastisch in ihrer Grose unterscheiden. In dieser Arbeit wird ein effektiver Algorithmus zur Verkleinerung der Matrixdarstellung vorgeschlagen, der die mit der Darstellung assoziierte Verteilung nicht verandert. Dieser Algorithmus verkleinert die Matrizen dabei betrachtlich starker als bereits existierende Methoden. Daruberhinaus bedient er sich nur numerischer Standardverfahren, wodurch er einfach zu implementieren ist. Wir identifizieren drei Operationen auf azyklischen Phasentypdarstellung, die in stochastischen Modellen haufig anzutreffen sind. Von diesen Operationen ausgehend entwickeln wir einen einfachen stochastischen Prozess-Kalkul, der eine grundlegende Struktur fur stochastische Modellierung und Analyse darstellt. Wir zeigen, dass die durch die drei Operationen erzeugten Darstellungen "beinahe gewiss" minimal sind und dass der Reduktionsalgorithmus benutzt werden kann, um diese beinahe gewiss minimalen Darstellungen zu erzeugen. Die Anwendbarkeit dieser Beitrage wird an einer Reihe von Fallstudien exemplifiziert.

[1]  Kishor S. Trivedi Probability and Statistics with Reliability, Queuing, and Computer Science Applications , 1984 .

[2]  Salvatore J. Bavuso,et al.  Dynamic fault-tree models for fault-tolerant computer systems , 1992 .

[3]  Boudewijn R. Haverkort,et al.  The Mean Value of the Maximum , 2002, PAPM-PROBMIV.

[4]  Marcel F. Neuts,et al.  Matrix-geometric solutions in stochastic models - an algorithmic approach , 1982 .

[5]  Peter Buchholz,et al.  A Novel Approach for Phase-Type Fitting with the EM Algorithm , 2006, IEEE Transactions on Dependable and Secure Computing.

[6]  Anthony Unwin,et al.  Reversibility and Stochastic Networks , 1980 .

[7]  Søren Asmussen,et al.  Phase-Type Representations in Random Walk and Queueing Problems , 1992 .

[8]  Mariëlle Stoelinga,et al.  A Compositional Semantics for Dynamic Fault Trees in Terms of Interactive Markov Chains , 2007, ATVA.

[9]  Joost-Pieter Katoen,et al.  A Markov reward model checker , 2005, Second International Conference on the Quantitative Evaluation of Systems (QEST'05).

[10]  Vaidyanathan Ramaswami,et al.  Introduction to Matrix Analytic Methods in Stochastic Modeling , 1999, ASA-SIAM Series on Statistics and Applied Mathematics.

[11]  Ramin Sadre,et al.  Fitting World Wide Web request traces with the EM-algorithim , 2001, SPIE ITCom.

[12]  Boudewijn R. Haverkort Performance of computer communication systems , 1998 .

[13]  Robert K. Brayton,et al.  Model-checking continuous-time Markov chains , 2000, TOCL.

[14]  Qi-Ming He,et al.  On matrix exponential distributions , 2007, Advances in Applied Probability.

[15]  Ching-tai Lin Structural controllability , 1974 .

[16]  William J. Stewart,et al.  Introduction to the numerical solution of Markov Chains , 1994 .

[17]  M. Dehon,et al.  A geometric interpretation of the relations between the exponential and generalized Erlang distributions , 1982, Advances in Applied Probability.

[18]  A. Cumani On the canonical representation of homogeneous markov processes modelling failure - time distributions , 1982 .

[19]  Christel Baier,et al.  Model-Checking Algorithms for Continuous-Time Markov Chains , 2002, IEEE Trans. Software Eng..

[20]  W. G. Marchal,et al.  Characterizations of generalized hyperexponential distribution functions , 1987 .

[21]  Colm Art O'Cinneide,et al.  Phase-type distributions and invariant polytopes , 1991, Advances in Applied Probability.

[22]  Holger Hermanns,et al.  On the use of MTBDDs for performability analysis and verification of stochastic systems , 2003, J. Log. Algebraic Methods Program..

[23]  C. Commault,et al.  Sparse representations of phase-type distributions , 1999 .

[24]  Carl D. Meyer,et al.  Matrix Analysis and Applied Linear Algebra , 2000 .

[25]  Robert E. Tarjan,et al.  Three Partition Refinement Algorithms , 1987, SIAM J. Comput..

[26]  Mario Bravetti Revisiting Interactive Markov Chains , 2002, Electron. Notes Theor. Comput. Sci..

[27]  Roberto Gorrieri,et al.  Extended Markovian Process Algebra , 1996, CONCUR.

[28]  陈永武 α , 1995 .

[29]  H. Tijms Understanding Probability: Chance Rules in Everyday Life , 2004 .

[30]  Mariëlle Stoelinga,et al.  Dynamic Fault Tree Analysis Using Input/Output Interactive Markov Chains , 2007, 37th Annual IEEE/IFIP International Conference on Dependable Systems and Networks (DSN'07).

[31]  R. Tyrrell Rockafellar,et al.  Convex Analysis , 1970, Princeton Landmarks in Mathematics and Physics.

[32]  David Assaf,et al.  Closure of Phase Type Distributions Under Operations Arising in Reliability Theory , 1982 .

[33]  Jane Hillston,et al.  An Efficient Kronecker Representation for PEPA Models , 2001, PAPM-PROBMIV.

[34]  Sheldon M. Ross,et al.  Introduction to probability models , 1975 .

[35]  Ludolf E. Meester,et al.  Stochastic delay propagation in railway networks and phase-type distributions , 2007 .

[36]  Tommaso Bolognesi,et al.  Tableau methods to describe strong bisimilarity on LOTOS processes involving pure interleaving and enabling , 1994, FORTE.

[37]  Markus Siegle,et al.  Recent Extensions to the Stochastic Process Algebra Tool CASPA , 2008, 2008 Fifth International Conference on Quantitative Evaluation of Systems.

[38]  H. Hermanns,et al.  Syntax , Semantics , Equivalences , and Axioms for MTIPP y , 1994 .

[39]  C. A. R. Hoare,et al.  Communicating sequential processes , 1978, CACM.

[40]  Gordon D. Plotkin,et al.  A structural approach to operational semantics , 2004, J. Log. Algebraic Methods Program..

[41]  Peter W. Glynn,et al.  Computing Poisson probabilities , 1988, CACM.

[42]  C. Loan,et al.  Nineteen Dubious Ways to Compute the Exponential of a Matrix , 1978 .

[43]  Ilaria Castellani,et al.  A non-interleaving semantics for CCS based on proved transitions , 1988 .

[44]  Holger Hermanns,et al.  Construction and Verification of Performance and Reliability Models , 2001, Bull. EATCS.

[45]  A. Jensen,et al.  Markoff chains as an aid in the study of Markoff processes , 1953 .

[46]  Jaynarayan H. Lala,et al.  FAULT-TOLERANT PARALLEL PROCESSOR , 1991 .

[47]  李幼升,et al.  Ph , 1989 .

[48]  Verena Wolf,et al.  Equivalences on Phase Type Processes , 2008 .

[49]  Bernhard Steffen,et al.  Reactive, generative, and stratified models of probabilistic processes , 1990, [1990] Proceedings. Fifth Annual IEEE Symposium on Logic in Computer Science.

[50]  Ernest J. Henley,et al.  Probabilistic risk assessment : reliability engineering, design, and analysis , 1992 .