Gaussian Markov Random Fields for Discrete Optimization via Simulation: Framework and Algorithms

This paper lays the foundation for employing Gaussian Markov random fields (GMRFs) for discrete decision–variable optimization via simulation; that is, optimizing the performance of a simulated system. Gaussian processes have gained popularity for inferential optimization, which iteratively updates a model of the simulated solutions and selects the next solution to simulate by relying on statistical inference from that model. We show that, for a discrete problem, GMRFs, a type of Gaussian process defined on a graph, provides better inference on the remaining optimality gap than the typical choice of continuous Gaussian process and thereby enables the algorithm to search efficiently and stop correctly when the remaining optimality gap is below a predefined threshold. We also introduce the concept of multiresolution GMRFs for large-scale problems, with which GMRFs of different resolutions interact to efficiently focus the search on promising regions of solutions.

[1]  Barry L. Nelson,et al.  Optimization via simulation over discrete decision variables , 2010 .

[2]  A. Winsor Sampling techniques. , 2000, Nursing times.

[3]  Barry L. Nelson,et al.  Stochastic kriging for simulation metamodeling , 2008, 2008 Winter Simulation Conference.

[4]  N. Zheng,et al.  Global Optimization of Stochastic Black-Box Systems via Sequential Kriging Meta-Models , 2006, J. Glob. Optim..

[5]  Aki Vehtari,et al.  Modelling local and global phenomena with sparse Gaussian processes , 2008, UAI.

[6]  Jie Xu,et al.  Industrial strength COMPASS: A comprehensive algorithm and software for optimization via simulation , 2010, TOMC.

[7]  Leonhard Held,et al.  Gaussian Markov Random Fields: Theory and Applications , 2005 .

[8]  K. Chen,et al.  Matrix preconditioning techniques and applications , 2005 .

[9]  Barry L. Nelson,et al.  A fully sequential procedure for indifference-zone selection in simulation , 2001, TOMC.

[10]  Barry L. Nelson,et al.  Discrete optimization via simulation using Gaussian Markov random fields , 2014, Proceedings of the Winter Simulation Conference 2014.

[11]  Sonja Kuhnt,et al.  Design and analysis of computer experiments , 2010 .

[12]  P. Frazier,et al.  Tutorial: Optimization via simulation with Bayesian statistics and dynamic programming , 2012, Proceedings Title: Proceedings of the 2012 Winter Simulation Conference (WSC).

[13]  Julien Bect,et al.  Robust Gaussian Process-Based Global Optimization Using a Fully Bayesian Expected Improvement Criterion , 2011, LION.

[14]  W. F. Tinney,et al.  On computing certain elements of the inverse of a sparse matrix , 1975, Commun. ACM.

[15]  Peter I. Frazier,et al.  Value of information methods for pairwise sampling with correlations , 2011, Proceedings of the 2011 Winter Simulation Conference (WSC).

[16]  Jun Yan,et al.  Gaussian Markov Random Fields: Theory and Applications , 2006 .

[17]  Yousef Saad,et al.  Iterative methods for sparse linear systems , 2003 .

[18]  K. Miller On the Inverse of the Sum of Matrices , 1981 .

[19]  พงศ์ศักดิ์ บินสมประสงค์,et al.  FORMATION OF A SPARSE BUS IMPEDANCE MATRIX AND ITS APPLICATION TO SHORT CIRCUIT STUDY , 1980 .

[20]  H. Niessner,et al.  On computing the inverse of a sparse matrix , 1983 .

[21]  M. Sasena,et al.  Exploration of Metamodeling Sampling Criteria for Constrained Global Optimization , 2002 .

[22]  P. Diggle,et al.  Geostatistical inference under preferential sampling , 2010 .

[23]  Thomas J. Santner,et al.  Sequential design of computer experiments to minimize integrated response functions , 2000 .

[24]  Jie Xu,et al.  Efficient discrete optimization via simulation using stochastic kriging , 2012, Proceedings Title: Proceedings of the 2012 Winter Simulation Conference (WSC).

[25]  David H. Wolpert,et al.  No free lunch theorems for optimization , 1997, IEEE Trans. Evol. Comput..

[26]  Loo Hay Lee,et al.  Simulation optimization via kriging: a sequential search using expected improvement with computing budget constraints , 2013 .

[27]  A. Law,et al.  A procedure for selecting a subset of size m containing the l best of k independent normal populations, with applications to simulation , 1985 .

[28]  Peter I. Frazier,et al.  Bayesian Optimization via Simulation with Pairwise Sampling and Correlated Prior Beliefs , 2016, Oper. Res..

[29]  Lihua Sun,et al.  Optimization via simulation using Gaussian Process-based Search , 2011, Proceedings of the 2011 Winter Simulation Conference (WSC).

[30]  Andy J. Keane,et al.  On the Design of Optimization Strategies Based on Global Response Surface Approximation Models , 2005, J. Glob. Optim..

[31]  Warren B. Powell,et al.  The Knowledge-Gradient Policy for Correlated Normal Beliefs , 2009, INFORMS J. Comput..

[32]  Victor Picheny,et al.  Comparison of Kriging-Based Methods for Simulation Optimization with Heterogeneous Noise , 2015 .

[33]  Donald R. Jones,et al.  Efficient Global Optimization of Expensive Black-Box Functions , 1998, J. Glob. Optim..

[34]  YANQING CHEN,et al.  Algorithm 8 xx : CHOLMOD , supernodal sparse Cholesky factorization and update / downdate ∗ , 2006 .