Multigrid with Rough Coefficients and Multiresolution Operator Decomposition from Hierarchical Information Games

Abstract

Abstract. We introduce a near-linear complexity (geometric and meshless/algebraic) multigrid/multiresolution method for PDEs with rough (L ∞ ) coefficients with rigorous a priori accuracy and performance estimates. The method is discovered through a decision/game theory formulation of the problems of (1) identifying restriction and interpolation operators, (2) recovering a signal from incomplete measurements based on norm constraints on its image under a linear operator, and (3) gambling on the value of the solution of the PDE based on a hierarchy of nested measurements of its solution or source term. The resulting elementary gambles form a hierarchy of (deterministic) basis functions of H 1 0 (Ω) (gamblets) that (1) are orthogonal across subscales/subbands with respect to the scalar product induced by the energy norm of the PDE, (2) enable sparse compression of the solution space in H 1 0 (Ω), and (3) induce an orthogonal multiresolution operator decomposition. The operating diagram of the multigrid method is that of an inverted pyramid in which gamblets are computed locally (by virtue of their exponential decay) and hierarchically (from fine to coarse scales) and the PDE is decomposed into a hierarchy of independent linear systems with uniformly bounded condition numbers. The resulting algorithm is parallelizable both in space (via localization) and in bandwidth/subscale (subscales can be computed independently from each other). Although the method is deterministic, it has a natural Bayesian interpretation under the measure of probability emerging (as a mixed strategy) from the information game formulation, and multiresolution approximations form a martingale with respect to the filtration induced by the hierarchy of nested measurements.

Constraint Energy Minimizing Generalized Multiscale Finite Element Method

“…It is in general not an easy task to derive a multiscale method with a convergence depends only on the coarse mesh size and independent of scales and contrast. To obtain multiscale methods with mesh dependent convergence, several approaches are considered in literature [39,32,38,22,10,6]. The theory of GMsFEM motivates the use of local spectral problems to capture the effects of high contrast channels.…”

Accelerated scale bridging with sparsely approximated Gaussian learning

“…The theoretical foundation of the hierarchical sparse Cholesky decomposition algorithm relies on the exponential localization property of a set of multi-resolution basis functions called gamblets [28,29]. We briefly go over the definition of gamblets and its important properties in this section.…”

A Model Reduction Method for Multiscale Elliptic Pdes with Random Coefficients Using an Optimization Approach

“…We consider the physical component first since the form (18) almost surely satisfies the constraint in the random space. For a deterministic elliptic PDE, it has been proven that ψ ·,k (x, ξ(ω * )) = g i,k (x)H k (ξ(ω * ))+R i,k (x, ξ(ω * ) will decay exponentially fast away from physical node x i [24,30,18]. Therefore, we know that ψ ·,k (x, ξ(ω)) has the exponential decay property in the physical space almost surely.…”