Preface | p. vii |
Background | |
Linear Algebra | p. 3 |
Vectors | p. 3 |
Matrices and Matrix Operations | p. 5 |
Matrices and Mappings | p. 6 |
Inverse and Generalized Inverse Matrices | p. 8 |
Direct Methods for Solving Linear Equations | p. 9 |
Norms | p. 12 |
Scalar Products | p. 14 |
Eigenvalues and Eigenvectors | p. 17 |
Matrix Decompositions | p. 19 |
Penalized Matrices | p. 22 |
Optimization | p. 27 |
Optimization Problems and Solutions | p. 27 |
Unconstrained Quadratic Programming | p. 28 |
Quadratic Cost Functions | p. 28 |
Unconstrained Minimization of Quadratic Functions | p. 29 |
Convexity | p. 31 |
Convex Quadratic Functions | p. 32 |
Local and Global Minimizers of Convex Function | p. 34 |
Existence of Minimizers | p. 35 |
Projections to Convex Sets | p. 36 |
Equality Constrained Problems | p. 38 |
Optimality Conditions | p. 39 |
Existence and Uniqueness | p. 41 |
KKT Systems | p. 42 |
Min-max, Dual, and Saddle Point Problems | p. 44 |
Sensitivity | p. 46 |
Error Analysis | p. 47 |
Inequality Constrained Problems | p. 49 |
Polyhedral Sets | p. 49 |
Farkas's Lemma | p. 50 |
Necessary Optimality Conditions for Local Solutions | p. 51 |
Existence and Uniqueness | p. 52 |
Optimality Conditions for Convex Problems | p. 54 |
Optimality Conditions for Bound Constrained Problems | p. 55 |
Min-max, Dual, and Saddle Point Problems | p. 55 |
Equality and Inequality Constrained Problems | p. 57 |
Optimality Conditions | p. 58 |
Existence and Uniqueness | p. 59 |
Partially Bound and Equality Constrained Problems | p. 59 |
Duality for Dependent Constraints | p. 61 |
Duality for Semicoercive Problems | p. 64 |
Linear Programming | p. 69 |
Solvability and Localization of Solutions | p. 69 |
Duality in Linear Programming | p. 70 |
Algorithms | |
Conjugate Gradients for Unconstrained Minimization | p. 73 |
Conjugate Directions and Minimization | p. 74 |
Generating Conjugate Directions and Krylov Spaces | p. 77 |
Conjugate Gradient Method | p. 78 |
Restarted CG and the Gradient Method | p. 81 |
Rate of Convergence and Optimality | p. 82 |
Min-max Estimate | p. 82 |
Estimate in the Condition Number | p. 84 |
Convergence Rate of the Gradient Method | p. 86 |
Optimality | p. 87 |
Preconditioned Conjugate Gradients | p. 87 |
Preconditioning by Conjugate Projector | p. 90 |
Conjugate Projectors | p. 90 |
Minimization in Subspace | p. 91 |
Conjugate Gradients in Conjugate Complement | p. 92 |
Preconditioning Effect | p. 94 |
Conjugate Gradients for More General Problems | p. 96 |
Convergence in Presence of Rounding Errors | p. 97 |
Numerical Experiments | p. 98 |
Basic CG and Preconditioning | p. 98 |
Numerical Demonstration of Optimality | p. 99 |
Comments and Conclusions | p. 100 |
Equality Constrained Minimization | p. 103 |
Review of Alternative Methods | p. 105 |
Penalty Method | p. 107 |
Minimization of Augmented Lagrangian | p. 108 |
An Optimal Feasibility Error Estimate for Homogeneous Constraints | p. 109 |
Approximation Error and Convergence | p. 111 |
Improved Feasibility Error Estimate | p. 112 |
Improved Approximation Error Estimate | p. 113 |
Preconditioning Preserving Gap in the Spectrum | p. 115 |
Exact Augmented Lagrangian Method | p. 116 |
Algorithm | p. 117 |
Convergence of Lagrange Multipliers | p. 119 |
Effect of the Steplength | p. 120 |
Convergence of the Feasibility Error | p. 124 |
Convergence of Primal Variables | p. 124 |
Implementation | p. 125 |
Asymptotically Exact Augmented Lagrangian Method | p. 126 |
Algorithm | p. 126 |
Auxiliary Estimates | p. 127 |
Convergence Analysis | p. 128 |
Adaptive Augmented Lagrangian Method | p. 130 |
Algorithm | p. 131 |
Convergence of Lagrange Multipliers for Large <$>\varrho<$> | p. 132 |
R-Linear Convergence for Any Initialization of <$>\varrho<$> | p. 134 |
Semimonotonic Augmented Lagrangians (SMALE) | p. 135 |
SMALE Algorithm | p. 136 |
Relations for Augmented Lagrangians | p. 137 |
Convergence and Monotonicity | p. 139 |
Linear Convergence for Large <$>\varrho_0<$> | p. 142 |
Optimality of the Outer Loop | p. 143 |
Optimality of SMALE with Conjugate Gradients | p. 145 |
Solution of More General Problems | p. 147 |
Implementation of Inexact Augmented Lagrangians | p. 148 |
Stopping, Modification of Constraints, and Preconditioning | p. 148 |
Initialization of Constants | p. 148 |
Numerical Experiments | p. 150 |
Uzawa, Exact Augmented Lagrangians, and SMALE | p. 150 |
Numerical Demonstration of Optimality | p. 151 |
Comments and References | p. 152 |
Bound Constrained Minimization | p. 155 |
Review of Alternative Methods | p. 157 |
KKT Conditions and Related Inequalities | p. 158 |
The Working Set Method with Exact Solutions | p. 160 |
Auxiliary Problems | p. 160 |
Algorithm | p. 161 |
Finite Termination | p. 164 |
Polyak's Algorithm | p. 165 |
Basic Algorithm | p. 165 |
Finite Termination | p. 166 |
Characteristics of Polyak's Algorithm | p. 167 |
Inexact Polyak's Algorithm | p. 167 |
Looking Ahead and Estimate | p. 167 |
Looking Ahead Polyak's Algorithm | p. 170 |
Easy Re-release Polyak's Algorithm | p. 171 |
Properties of Modified Polyak's Algorithms | p. 172 |
Gradient Projection Method | p. 173 |
Conjugate Gradient Versus Gradient Projections | p. 174 |
Contraction in the Euclidean Norm | p. 175 |
The Fixed Steplength Gradient Projection Method | p. 177 |
Quadratic Functions with Identity Hessian | p. 178 |
Dominating Function and Decrease of the Cost Function | p. 181 |
Modified Proportioning with Gradient Projections | p. 184 |
MPGP Schema | p. 184 |
Rate of Convergence | p. 186 |
Modified Proportioning with Reduced Gradient Projections | p. 189 |
MPRGP Schema | p. 189 |
Rate of Convergence | p. 190 |
Rate of Convergence of Projected Gradient | p. 193 |
Optimality | p. 197 |
Identification Lemma and Finite Termination | p. 198 |
Finite Termination for Dual Degenerate Solution | p. 201 |
Implementation of MPRGP with Optional Modifications | p. 204 |
Expansion Step with Feasible Half-Step | p. 204 |
MPRGP Algorithm | p. 205 |
Unfeasible MPRGP | p. 206 |
Choice of Parameters | p. 208 |
Dynamic Release Coefficient | p. 209 |
Preconditioning | p. 210 |
Preconditioning in Face | p. 210 |
Preconditioning by Conjugate Projector | p. 212 |
Numerical Experiments | p. 216 |
Polyak, MPRGP, and Preconditioned MPRGP | p. 216 |
Numerical Demonstration of Optimality | p. 217 |
Comments and References | p. 218 |
Bound and Equality Constrained Minimization | p. 221 |
Review of the Methods for Bound and Equality Constrained Problems | p. 222 |
SMALBE Algorithm for Bound and Equality Constraints | p. 223 |
KKT Conditions and Projected Gradient | p. 223 |
SMALBE Algorithm | p. 223 |
Inequalities Involving the Augmented Lagrangian | p. 225 |
Monotonicity and Feasibility | p. 227 |
Boundedness | p. 229 |
Convergence | p. 233 |
Optimality of the Outer Loop | p. 235 |
Optimality of the Inner Loop | p. 237 |
Solution of More General Problems | p. 239 |
Implementation | p. 240 |
SMALBE-M | p. 241 |
Numerical Experiments | p. 242 |
Balanced Reduction of Feasibility and Gradient Errors | p. 242 |
Numerical Demonstration of Optimality | p. 243 |
Comments and References | p. 244 |
Applications to Variational Inequalities | |
Solution of a Coercive Variational Inequality by FETI-DP Method | p. 249 |
Model Coercive Variational Inequality | p. 250 |
FETI-DP Domain Decomposition and Discretization | p. 251 |
Optimality | p. 254 |
Numerical Experiments | p. 255 |
Comments and References | p. 256 |
Solution of a Semicoercive Variational Inequality by TFETI Method | p. 259 |
Model Semicoercive Variational Inequality | p. 260 |
TFETI Domain Decomposition and Discretization | p. 261 |
Natural Coarse Grid | p. 264 |
Optimality | p. 265 |
Numerical Experiments | p. 267 |
Comments and References | p. 269 |
References | p. 271 |
Index | p. 281 |
Table of Contents provided by Ingram. All Rights Reserved. |