Approximate aggregation of Markovian models using alternating least squares

State based analysis of Markovian models is faced with the problem of state space explosion. To handle huge state spaces often compositional modeling and aggregation of components are used. Exact aggregation resulting in exact transient or stationary results is only possible in some cases, when the Markov process is lumpable. Therefore approximate aggregation is often applied to reduce the state space. Several approximate aggregation methods exist which are usually based on heuristics. This paper presents a new aggregation approach for Markovian components which computes aggregates that minimize the difference according to some algebraically defined function which describes the difference between the component and the aggregate. If the difference becomes zero, aggregation is exact, which means that component and aggregate are indistinguishable in the sense that transient and stationary results in any environment are identical. For the computation of aggregates, an alternating least squares approach is presented which tries to minimize the norm-wise difference between the original component and the aggregate. Algorithms to compute aggregates are also introduced and the quality of the approximation is evaluated by means of several examples.

[1]  Wim P. Krijnen,et al.  Convergence of the sequence of parameters generated by alternating least squares algorithms , 2006, Comput. Stat. Data Anal..

[2]  Yu-Jin Zhang,et al.  Nonnegative Matrix Factorization: A Comprehensive Review , 2013, IEEE Transactions on Knowledge and Data Engineering.

[3]  P. Buchholz,et al.  Complexity of Kronecker Operations on Sparse Matrices with Applications to the Solution of Markov Models , 1997 .

[4]  Amy Nicole Langville,et al.  Algorithms, Initializations, and Convergence for the Nonnegative Matrix Factorization , 2014, ArXiv.

[5]  G. Golub,et al.  A Hessenberg-Schur method for the problem AX + XB= C , 1979 .

[6]  William H. Sanders,et al.  Bound-Preserving Composition for Markov Reward Models , 2006, Third International Conference on the Quantitative Evaluation of Systems - (QEST'06).

[7]  J. Hillston.  Compositional Markovian Modelling Using a Process Algebra , 1995 .

[8]  Brigitte Plateau.  On the stochastic structure of parallelism and synchronization models for distributed algorithms , 1985, SIGMETRICS 1985.

[9]  John G. Kemeny,et al.  Finite Markov chains , 1960 .

[10]  P.-J. Courtois,et al.  Decomposability: Queueing and Computer System Applications , 2014 .

[11]  J. Leeuw,et al.  Principal component analysis of three-mode data by means of alternating least squares algorithms , 1980 .

[12]  Mogens Bladt,et al.  Point processes with finite-dimensional conditional probabilities , 1999 .

[13]  P. Buchholz.  Exact and ordinary lumpability in finite Markov chains , 1994, Journal of Applied Probability.

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

[15]  Stephen Gilmore,et al.  Explorer Compositional Approximate Markov Chain Aggregation for PEPA Models , 2016 .

[16]  Ana Busic,et al.  Bounds for Point and Steady-State Availability: An Algorithmic Approach Based on Lumpability and Stochastic Ordering , 2005, EPEW/WS-FM.

[17]  Donghui Chen,et al.  Nonnegativity constraints in numerical analysis , 2009, The Birth of Numerical Analysis.

[18]  Peter Buchholz,et al.  A class of hierarchical queueing networks and their analysis , 1994, Queueing Syst. Theory Appl..

[19]  G. Franceschinis,et al.  Bounds for Quasi-Lumpable Markow Chains , 1994, Perform. Evaluation.

[20]  Evgenia Smirni,et al.  KPC-Toolbox: Simple Yet Effective Trace Fitting Using Markovian Arrival Processes , 2008, 2008 Fifth International Conference on Quantitative Evaluation of Systems.

[21]  Peter Buchholz.  Lumpability and nearly-lumpability in hierarchical queueing networks , 1995, Proceedings of 1995 IEEE International Computer Performance and Dependability Symposium.

[22]  D. Sorensen,et al.  Approximation of large-scale dynamical systems: an overview , 2004 .

[23]  Peter Buchholz,et al.  Composition and Equivalence of Markovian and Non-Markovian Models , 2011, 2011 Eighth International Conference on Quantitative Evaluation of SysTems.

[24]  Michael Piatek,et al.  Tsnnls: A solver for large sparse least squares problems with non-negative variables , 2004, ArXiv.

[25]  Peter Buchholz,et al.  Rational Processes Related to Communicating Markov Processes , 2012, Journal of Applied Probability.

[26]  Stephen A. Vavasis,et al.  On the Complexity of Nonnegative Matrix Factorization , 2007, SIAM J. Optim..

[27]  Holger Hermanns,et al.  Interactive Markov Chains , 2002, Lecture Notes in Computer Science.

[28]  Gene H. Golub,et al.  Matrix computations , 1983 .

[29]  P. Buchholz.  Equivalence Relations for Stochastic Automata Networks , 1995 .

[30]  Jean-Michel Fourneau,et al.  A Methodology for Solving Markov Models of Parallel Systems , 1991, J. Parallel Distributed Comput..

[31]  Sally Floyd,et al.  Wide area traffic: the failure of Poisson modeling , 1995, TNET.

[32]  Walter Willinger,et al.  On the self-similar nature of Ethernet traffic , 1993, SIGCOMM '93.

[33]  Scott A. Smolka,et al.  Composition and Behaviors of Probabilistic I/O Automata , 1994, Theor. Comput. Sci..

[34]  Karen H. Haskell,et al.  An algorithm for linear least squares problems with equality and nonnegativity constraints , 1981, Math. Program..

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

[36]  Susanna Donatelli.  Superposed stochastic automata: a class of stochastic Petri nets amenable to parallel solution , 1991, Proceedings of the Fourth International Workshop on Petri Nets and Performance Models PNPM91.

[37]  Peter Buchholz,et al.  On minimal representations of Rational Arrival Processes , 2013, Ann. Oper. Res..

[38]  Peter Buchholz,et al.  Aggregation of Markovian Models -- An Alternating Least Squares Approach , 2012, 2012 Ninth International Conference on Quantitative Evaluation of Systems.

[39]  Peter Buchholz,et al.  Complexity of Memory-Efficient Kronecker Operations with Applications to the Solution of Markov Models , 2000, INFORMS J. Comput..

[40]  Michael W. Berry,et al.  Algorithms and applications for approximate nonnegative matrix factorization , 2007, Comput. Stat. Data Anal..

[41]  Susanna Donatelli,et al.  Superposed Stochastic Automata: A Class of Stochastic Petri Nets with Parallel Solution and Distributed State Space , 1993, Perform. Evaluation.

[42]  Gábor Horváth,et al.  A minimal representation of Markov arrival processes and a moments matching method , 2007, Perform. Evaluation.

[43]  Peter Buchholz,et al.  Stochastic Petri nets with matrix exponentially distributed firing times , 2010, Perform. Evaluation.

[44]  Nicolas Gillis,et al.  Using underapproximations for sparse nonnegative matrix factorization , 2009, Pattern Recognit..

[45]  Peter Buchholz.  Exact Performance Equivalence: An Equivalence Relation for Stochastic Automata , 1999, Theor. Comput. Sci..