| Coefficients of Variations in Analysis of Macro-policy Effects: An Example of Two-Parameter Poisson-Dirichlet Distributions | p. 1 |
| Introduction | p. 1 |
| The Model | p. 2 |
| Asymptotic Properties of the Number of Sectors | p. 2 |
| The Coefficients of Variation | p. 3 |
| The Number of Sectors | p. 3 |
| The Number of Sectors of Specified Size | p. 3 |
| Discussion | p. 3 |
| References | p. 4 |
| How Many Experiments Are Needed to Adapt? | p. 5 |
| Introduction | p. 5 |
| Worst-Case Approach to Adaptation | p. 7 |
| Worst-Case Performance | p. 7 |
| Adaptive Design | p. 8 |
| The Experimental Effort Needed for Adaptation | p. 9 |
| A Numerical Example | p. 11 |
| Conclusions | p. 13 |
| References | p. 14 |
| A Mutual Information Based Distance for Multivariate Gaussian Processes | p. 15 |
| Introduction | p. 15 |
| Model Class | p. 17 |
| Principal Angles, Canonical Correlations and Mutual Information | p. 19 |
| Principal Angles and Directions | p. 20 |
| Canonical Correlations | p. 20 |
| Mutual Information | p. 21 |
| Application to Stochastic Processes | p. 21 |
| A Distance Between Multivariate Gaussian Processes | p. 25 |
| Definition and Metric Properties | p. 25 |
| Computation | p. 26 |
| Special Case of Scalar Processes | p. 26 |
| Relation with Subspace Angles Between Scalar Stochastic Processes | p. 27 |
| Relation with a Cepstral Distance | p. 27 |
| The Cepstral Nature of the Mutual Information Distance | p. 28 |
| Multivariate Power Cepstrum and Cepstral Distance | p. 28 |
| The Cepstral Nature of the Mutual Information Distance | p. 29 |
| Conclusions and Open Problems | p. 30 |
| Conclusions | p. 30 |
| Open Problems | p. 31 |
| References | p. 31 |
| Differential Forms and Dynamical Systems | p. 35 |
| Introduction | p. 35 |
| Planar Dynamical Systems | p. 36 |
| The Principle of the Torus for Autonomous Systems | p. 40 |
| Lyapunov-Like Differential Forms for the Existence of Cross Sections | p. 41 |
| Necessary and Sufficient Conditions for Existence of Periodic Orbits | p. 42 |
| Stability and Robustness of Periodic Orbits | p. 43 |
| References | p. 44 |
| An Algebraic Framework for Bayes Nets of Time Series | p. 45 |
| Introduction | p. 45 |
| Conditional Independence and Stochastic Realization | p. 46 |
| Lattice Conditionally Independence and Stochastic Realization | p. 48 |
| Lattices of Subspaces | p. 48 |
| Lattice Conditionally Orthogonal Stochastic Hilbert Spaces | p. 49 |
| Spatially Patterned Infinite Bayes Nets | p. 53 |
| References | p. 56 |
| A Birds Eye View on System Identification | p. 59 |
| Introduction | p. 59 |
| Structure Theory | p. 61 |
| Estimation for a Given Subclass | p. 64 |
| Model Selection | p. 67 |
| Linear Non-mainstream Cases | p. 68 |
| Nonlinear Systems | p. 69 |
| Present State and Future Developments | p. 69 |
| References | p. 70 |
| Further Results on the Byrnes-Georgiou-Lindquist Generalized Moment Problem | p. 73 |
| Introduction | p. 73 |
| A Generalized Moment Problem | p. 74 |
| Kullback-Leibler Criterion | p. 75 |
| Optimality Conditions and the Dual Problem | p. 76 |
| An Existence Theorem | p. 77 |
| A Descent Method for the Dual Problem | p. 79 |
| References | p. 81 |
| Factor Analysis and Alternating Minimization | p. 85 |
| Introduction | p. 85 |
| The Model | p. 86 |
| Lifting of the Original Problem | p. 87 |
| The First Partial Minimization Problem | p. 88 |
| The Second Partial Minimization Problem | p. 89 |
| The Link to the Original Problem | p. 91 |
| Alternating Minimization Algorithm | p. 92 |
| The Algorithm | p. 92 |
| Proof of Proposition 1 | p. 94 |
| References | p. 95 |
| Appendix | p. 95 |
| Multivariate Normal Distribution | p. 95 |
| Partitioned Matrices | p. 95 |
| Tensored Polynomial Models | p. 97 |
| Introduction | p. 97 |
| Tensored Models | p. 98 |
| Preliminaries | p. 98 |
| Tensored Polynomial and Rational Models | p. 99 |
| Module Structures on Tensored Models | p. 101 |
| Duality | p. 103 |
| Homomorphisms of Tensored Models | p. 104 |
| Applications | p. 105 |
| The Space of Intertwining Maps | p. 105 |
| The Polynomial Sylvester Equation | p. 106 |
| Solving the Sylvester Equation | p. 107 |
| Invariant Factors of the Sylvester Map | p. 109 |
| References | p. 112 |
| Distances Between Time-Series and Their Autocorrelation Statistics | p. 113 |
| Introduction | p. 113 |
| Interpretation of the L[subscript 1] Distance | p. 114 |
| A Distance Between Covariance Matrices | p. 115 |
| An Example | p. 118 |
| Approximating Sample Covariances | p. 120 |
| Comparison with the von Neumann Entropy | p. 120 |
| Structured Covariances | p. 121 |
| References | p. 122 |
| Global Identifiability of Complex Models, Constructed from Simple Submodels | p. 123 |
| Introduction | p. 123 |
| The Problem | p. 124 |
| A Simple Example of Interconnected Modules | p. 126 |
| Preliminary Considerations and Tools | p. 127 |
| Identifiability Analysis | p. 128 |
| A Formal Theorem on Identifiability from Sub-models | p. 131 |
| Global Identifiability | p. 131 |
| Local Identifiability | p. 132 |
| Conclusions | p. 132 |
| References | p. 133 |
| Identification of Hidden Markov Models - Uniform LLN-s | p. 135 |
| Introduction | p. 135 |
| Hidden Markov Models | p. 136 |
| L-Mixing Processes | p. 138 |
| Asymptotic Properties of the Log-Likelihood Function | p. 139 |
| The Case of Primitive Q-s | p. 141 |
| The Derivative of the Predictive Filter | p. 144 |
| Uniform Laws of Large Numbers | p. 147 |
| Estimation of Hidden Markov Models | p. 149 |
| References | p. 149 |
| Identifiability and Informative Experiments in Open and Closed-Loop Identification | p. 151 |
| Introduction | p. 151 |
| The Prediction Error Identification Setup | p. 153 |
| Identifiability, Informative Data, and All That Jazz | p. 154 |
| Analysis of the Information Matrix | p. 157 |
| Expressions of the Pseudoregression Vector | p. 157 |
| The Range and Kernel of Rank-One Vector Processes | p. 158 |
| Regularity Conditions for I([theta]): A First Analysis | p. 159 |
| Rich and Exciting Signals | p. 161 |
| Regularity of I([theta]) for ARMAX and BJ Model Structures | p. 166 |
| Conclusions | p. 169 |
| References | p. 169 |
| On Interpolation and the Kimura-Georgiou Parametrization | p. 171 |
| Introduction | p. 171 |
| Interpolation Conditions as Matrix Equations | p. 172 |
| The Connection with the Kimura-Georgiou Parametrization | p. 177 |
| References | p. 182 |
| The Control of Error in Numerical Methods | p. 183 |
| Introduction | p. 183 |
| A Simple Example | p. 184 |
| Four-Step Adams-Bashforth | p. 186 |
| Statistical Analysis | p. 190 |
| Conclusion | p. 192 |
| References | p. 192 |
| Contour Reconstruction and Matching Using Recursive Smoothing Splines | p. 193 |
| Introduction | p. 193 |
| Problem Formulation and Motivation | p. 194 |
| Some Theoretical Properties | p. 196 |
| Proper Periodicity Conditions | p. 196 |
| Continuous Time, Continuous Data | p. 197 |
| Continuous Time, Discrete Data | p. 199 |
| Continuous Time, Discrete Data Iterated | p. 200 |
| Data Set Reconstruction | p. 200 |
| Evaluation of Recursive Spline Method | p. 202 |
| Conclusions | p. 205 |
| References | p. 206 |
| Role of LQ Decomposition in Subspace Identification Methods | p. 207 |
| Introduction | p. 207 |
| State-Input-Output Matrix Equation | p. 208 |
| MOESP Method | p. 209 |
| N4SID Method | p. 211 |
| Zero-Input Responses | p. 211 |
| Relation to Ho-Kalman's Method | p. 214 |
| State Vector and Zero-State Response | p. 214 |
| Zero-State Response | p. 215 |
| Conclusions | p. 216 |
| References | p. 217 |
| Canonical Operators on Graphs | p. 221 |
| Introduction | p. 221 |
| Graphs | p. 222 |
| The Geometry of Graphs and Digraphs | p. 222 |
| Operator Theory on Graphs and Digraphs | p. 224 |
| Differences, Divergences, Laplacians and Dirac Operators | p. 227 |
| Operators on Weighted Graphs | p. 230 |
| The Incidence Operator and Its Kin | p. 232 |
| The Drift of a Digraph | p. 235 |
| References | p. 236 |
| Prediction-Error Approximation by Convex Optimization | p. 239 |
| Introduction | p. 239 |
| Prediction-Error Approximation | p. 240 |
| Prediction-Error Approximation in Restricted Model Classes | p. 241 |
| The Kullback-Leibler Criterion and Maximum-Likelihood Identification | p. 245 |
| Prediction-Error Approximation by Analytic Interpolation | p. 246 |
| Conclusion | p. 248 |
| References | p. 248 |
| Patchy Solutions of Hamilton-Jacobi-Bellman Partial Differential Equations | p. 251 |
| Hamilton Jacobi Bellman PDEs | p. 251 |
| Other Approaches | p. 255 |
| New Approach | p. 255 |
| One Dimensional HJB PDEs | p. 256 |
| One Dimensional Example | p. 260 |
| HJB PDEs in Higher Dimensions | p. 261 |
| Two Dimensional Example | p. 268 |
| Conclusion | p. 269 |
| References | p. 269 |
| A Geometric Assignment Problem for Robotic Networks | p. 271 |
| Introduction | p. 271 |
| Geometric and Stochastic Preliminaries | p. 272 |
| The Euclidean Traveling Salesperson Problem | p. 273 |
| Bins and Balls | p. 273 |
| Random Geometric Graphs | p. 273 |
| Network Model and Problem Statement | p. 274 |
| Robotic Network Model | p. 274 |
| The Target Assignment Problem | p. 274 |
| Sparse and Dense Environments | p. 275 |
| Sparse Environments | p. 275 |
| Assignment-Based Algorithms with Lower Bound Analysis | p. 275 |
| The ETSP AssgmtAlgorithm with Upper Bound Analysis | p. 276 |
| Dense Environments | p. 279 |
| The Grid AssgmtAlgorithm with Complexity Analysis | p. 279 |
| A Sensor Based Version | p. 282 |
| Congestion Issues | p. 282 |
| Conclusion and Extensions | p. 283 |
| References | p. 283 |
| On the Distance Between Non-stationary Time Series | p. 285 |
| Introduction | p. 285 |
| Formalization | p. 286 |
| Introducing Nuisances | p. 287 |
| Dynamic Time Warping | p. 288 |
| Dynamics, or Lack Thereof, in DTW | p. 289 |
| Time Warping Under Dynamic Constraints | p. 290 |
| Going Blind | p. 292 |
| Computing the Distance | p. 293 |
| Correlation Kernels for Non-stationary Time Series | p. 294 |
| Invariance Via Canonization | p. 295 |
| Discussion | p. 297 |
| References | p. 298 |
| Stochastic Realization for Stochastic Control with Partial Observations | p. 301 |
| Introduction | p. 301 |
| Problem Formulation | p. 302 |
| The Classical Approach | p. 303 |
| The Stochastic Realization Approach to Stochastic Control with Partial Observations | p. 305 |
| Special Cases | p. 306 |
| Concluding Remarks | p. 312 |
| References | p. 312 |
| Experiences from Subspace System Identification - Comments from Process Industry Users and Researchers | p. 315 |
| Introduction | p. 315 |
| Questions and Answers from the User | p. 317 |
| Comments from the Researchers | p. 320 |
| Input Design | p. 321 |
| Merging Data | p. 322 |
| Merging of Models | p. 324 |
| Conclusion | p. 326 |
| References | p. 326 |
| Recursive Computation of the MPUM | p. 329 |
| Introduction | p. 329 |
| Problem Statement | p. 330 |
| Subspace Identification | p. 333 |
| State Construction by Past/Future Partition | p. 334 |
| The Hankel Structure and the Past/Future Partition | p. 336 |
| The Left Kernel of a Hankel Matrix | p. 338 |
| Recursive Computation of a Module Basis | p. 339 |
| Concluding Remarks | p. 341 |
| Subspace ID | p. 341 |
| State Construction by Shift-and-Cut | p. 342 |
| Return to the Data | p. 343 |
| Approximation and Balanced Reduction | p. 343 |
| The Complementary System | p. 343 |
| References | p. 344 |
| New Development of Digital Signal Processing Via Sampled-Data Control Theory | p. 345 |
| Foreword | p. 345 |
| Introduction | p. 345 |
| The Shannon Paradigm | p. 346 |
| Problems in the Shannon Paradigm | p. 347 |
| Control Theoretic Formulation | p. 349 |
| Application to Images | p. 352 |
| Concluding Remarks and Related Work | p. 354 |
| References | p. 355 |
| Table of Contents provided by Ingram. All Rights Reserved. |