Finite-dimensional Approximation Research Articles

Convex Approach to Data-Driven Off-Road Navigation via Linear Transfer Operators

We consider the problem of optimal control design for navigation on off-road terrain. We use a traversability measure to characterize the difficulty of navigation on off-road terrain. The traversability measure captures terrain properties essential for navigation, such as elevation maps, roughness, slope, and texture. The terrain with the presence or absence of obstacles becomes a particular case of the proposed traversability measure. We provide a convex formulation to the off-road navigation problem by lifting the problem to the density space using the linear Perron-Frobenius (P-F) operator. The convex formulation leads to an infinite-dimensional optimal navigation problem for control synthesis. We construct the finite-dimensional approximation of the optimization problem using data. We use a computational framework based on the data-driven approximation of the Koopman operator. This makes the proposed approach data-driven and applicable to cases where an explicit system model is unavailable. Finally, we apply the proposed navigation framework with single integrator dynamics and Dubin's car model.

Extending the extended dynamic mode decomposition with latent observables: the latent EDMD framework

Bernard O Koopman proposed an alternative view of dynamical systems based on linear operator theory, in which the time evolution of a dynamical system is analogous to the linear propagation of an infinite-dimensional vector of observables. In the last few years, several works have shown that finite-dimensional approximations of this operator can be extremely useful for several applications, such as prediction, control, and data assimilation. In particular, a Koopman representation of a dynamical system with a finite number of dimensions will avoid all the problems caused by nonlinearity in classical state-space models. In this work, the identification of finite-dimensional approximations of the Koopman operator and its associated observables is expressed through the inversion of an unknown augmented linear dynamical system. The proposed framework can be regarded as an extended dynamical mode decomposition that uses a collection of latent observables. The use of a latent dictionary applies to a large class of dynamical regimes, and it provides new means for deriving appropriate finite-dimensional linear approximations to high-dimensional nonlinear systems.

Data Assimilation for the Two-Dimensional Ambipolar Diffusion Equation in Earth’s Ionosphere Model

The problem of variational data assimilation for the INM RAS two-dimensional diffusion model of the Earth’s ionosphere F region is considered. Total integral electron contents along given paths are used as observation data. The general statement of the problem in differential form is formulated, and its solvability is analyzed. Based on a regularized statement, an iterative algorithm for solving the assimilation problem is constructed, and its convergence is demonstrated. A finite-dimensional approximation is constructed, the numerical solution of the problem is implemented, and the stability and convergence of the difference scheme are proved. The quality of the reconstruction of electron concentration fields is examined in test numerical experiments. It is shown that a weakly perturbed solution is reconstructed with acceptable accuracy for both stationary and evolutionary statements in the case of vertical and slant integration paths.

Finite Dimensional Approximations of Hamilton–Jacobi–Bellman Equations for Stochastic Particle Systems with Common Noise

Finite Dimensional Approximations of Hamilton–Jacobi–Bellman Equations for Stochastic Particle Systems with Common Noise

Finite dimensional approximations in operator algebras

A non-self-adjoint operator algebra is said to be residually finite dimensional (RFD) if it embeds into a product of matrix algebras. We characterize RFD operator algebras in terms of their matrix state space, and moreover show that an operator algebra is RFD if and only if every representation can be approximated by finite dimensional ones in the point weak operator topology. This is a non-self-adjoint version of a theorem of Exel and Loring for C⁎-algebras. Moreover, we construct an example of an operator algebra for which approximation in the point strong operator topology is not possible. As a consequence, the maximal C⁎-algebra generated by this operator algebra is not RFD. This answers questions of Clouâtre and Ramsey and of Clouâtre and Dor-On.

A Numerical Energy Reduction Approach for Semilinear Diffusion-Reaction Boundary Value Problems Based on Steady-State Iterations

We present a novel energy-based numerical analysis of semilinear diffusion-reaction boundary value problems, where the nonlinear reaction terms need to be neither monotone nor convex. Based on a suitable variational setting, the proposed computational scheme can be seen as an energy reduction approach. More specifically, this procedure aims to generate a sequence of numerical approximations, which results from the iterative solution of related (stabilized) linearized discrete problems, and tends to a critical point of the underlying energy functional in a stable way. Simultaneously, the finite-dimensional approximation spaces are adaptively refined. This is implemented in terms of a new mesh refinement strategy in the context of finite element discretizations, which again relies on the energy structure of the problem under consideration. In particular, in contrast to more traditional approaches, it does not involve any a posteriori error estimators, and is based on local energy reduction indicators instead. In combination, the resulting adaptive algorithm consists of an iterative linearization procedure on a sequence of hierarchically refined discrete spaces, which we prove to converge toward a solution of the continuous problem in an appropriate sense. Numerical experiments demonstrate the robustness and reliability of our approach for a series of examples.

Finite-Dimensional Approximations and Semigroup Coactions for Operator Algebras

Abstract The residual finite-dimensionality of a ${\textrm{C}}^{\ast }$-algebra is known to be encoded in a topological property of its space of representations, stating that finite-dimensional representations should be dense therein. We extend this paradigm to general (possibly non-self-adjoint) operator algebras. While numerous subtleties emerge in this greater generality, we exhibit novel tools for constructing finite-dimensional approximations. One such tool is a notion of a residually finite-dimensional coaction of a semigroup on an operator algebra, which allows us to construct finite-dimensional approximations for operator algebras of functions and operator algebras of semigroups. Our investigation is intimately related to the question of when residual finite-dimensionality of an operator algebra is inherited by its maximal ${\textrm{C}}^{\ast }$-cover, which we establish in many cases of interest.

The Strong Converse Exponent of Discriminating Infinite-Dimensional Quantum States

The sandwiched Rényi divergences of two finite-dimensional density operators quantify their asymptotic distinguishability in the strong converse domain. This establishes the sandwiched Rényi divergences as the operationally relevant ones among the infinitely many quantum extensions of the classical Rényi divergences for Rényi parameter alpha >1. The known proof of this goes by showing that the sandwiched Rényi divergence coincides with the regularized measured Rényi divergence, which in turn is proved by asymptotic pinching, a fundamentally finite-dimensional technique. Thus, while the notion of the sandwiched Rényi divergences was extended recently to density operators on an infinite-dimensional Hilbert space (in fact, even for states of an arbitrary von Neumann algebra), these quantities were so far lacking an operational interpretation similar to the finite-dimensional case, and it has also been open whether they coincide with the regularized measured Rényi divergences. In this paper we fill this gap by answering both questions in the positive for density operators on an infinite-dimensional Hilbert space, using a simple finite-dimensional approximation technique. We also initiate the study of the sandwiched Rényi divergences, and the related problem of the strong converse exponent, for pairs of positive semi-definite operators that are not necessarily trace-class (this corresponds to considering weights in a general von Neumann algebra setting). This is motivated by the need to define conditional Rényi entropies in the infinite-dimensional setting, while it might also be interesting from the purely mathematical point of view of extending the concept of Rényi (and other) divergences to settings beyond the standard one of positive trace-class operators (positive normal functionals in the von Neumann algebra setting). In this spirit, we also discuss the definition and some properties of the more general family of Rényi (alpha ,z)-divergences of positive semi-definite operators on an infinite-dimensional separable Hilbert space.

Integration of Physics- and Data-Driven Power System Models in Transient Analysis After Major Disturbances

The article explores the analysis of transient phenomena in large-scale power systems subjected to major disturbances from the aspect of interleaving, coordinating, and refining physics- and data-driven models. Major disturbances can lead to cascading failures and ultimately to the partial power system blackout. Our primary interest is in a framework that would enable coordinated and seamlessly integrated use of the two types of models in engineered systems. Parts of this framework include: 1) optimized compressed sensing, 2) customized finite-dimensional approximations of the Koopman operator, and 3) gray-box integration of physics-driven (equation-based) and data-driven (deep neural network-based) models. The proposed three-stage procedure is applied to the transient stability analysis on the multimachine benchmark example of a 441-bus real-world test system, where the results are shown for a synchronous generator with local measurements in the connection point.

Quasi-Invariance for Infinite-Dimensional Kolmogorov Diffusions

We prove Cameron-Martin type quasi-invariance results for the heat kernel measure of infinite-dimensional Kolmogorov and related diffusions. We first study quantitative functional inequalities for appropriate finite-dimensional approximations of these diffusions, and we prove these inequalities hold with dimension-independent coefficients. Applying an approach developed by Baudoin, Driver, Gordina, and Melcher previously, these uniform bounds may then be used to prove that the heat kernel measure for certain of these infinite-dimensional diffusions is quasi-invariant under changes of the initial state.

A new algorithm for computing path integrals and weak approximation of SDEs inspired by large deviations and Malliavin calculus

The paper gives a novel path integral formula inspired by large deviation theory and Malliavin calculus. The proposed finite-dimensional approximation of integrals on path space will be a new higher-order weak approximation of multidimensional stochastic differential equations where the dominant part of the local expansion is governed by Varadhan's geodesic distance and the correction terms are given as Malliavin weights. An optimal truncation of asymptotic expansion is used to reduce computational effort. Kusuoka's estimate is applied to justify the finite-dimensional approximation of path integrals. An efficient simulation method is provided with the algorithm. Numerical results are shown to verify the effectiveness.

A non-stationary iterative Tikhonov regularization method for simultaneous inversion in a time-fractional diffusion equation

The present paper is devoted to recovering the source term and the initial data simultaneously for a time-fractional diffusion equation from additional temperature data at two fixed times t=T1 and t=T2. Firstly, we prove the uniqueness result for the direct problem. And then we apply a non-stationary iterative Tikhonov regularization method to solve the inverse problem and propose a finite dimensional approximation algorithm. Three numerical examples in both one-dimensional and two-dimensional cases are provided to demonstrate the effectiveness and feasibility of the proposed method.

Computation of nonparametric, mixed effects, maximum likelihood, biosensor data based-estimators for the distributions of random parameters in an abstract parabolic model for the transdermal transport of alcohol.

The existence and consistency of a maximum likelihood estimator for the joint probability distribution of random parameters in discrete-time abstract parabolic systems was established by taking a nonparametric approach in the context of a mixed effects statistical model using a Prohorov metric framework on a set of feasible measures. A theoretical convergence result for a finite dimensional approximation scheme for computing the maximum likelihood estimator was also established and the efficacy of the approach was demonstrated by applying the scheme to the transdermal transport of alcohol modeled by a random parabolic partial differential equation (PDE). Numerical studies included show that the maximum likelihood estimator is statistically consistent, demonstrated by the convergence of the estimated distribution to the "true" distribution in an example involving simulated data. The algorithm developed was then applied to two datasets collected using two different transdermal alcohol biosensors. Using the leave-one-out cross-validation (LOOCV) method, we found an estimate for the distribution of the random parameters based on a training set. The input from a test drinking episode was then used to quantify the uncertainty propagated from the random parameters to the output of the model in the form of a 95 error band surrounding the estimated output signal.

Non-Stationary Dynamic Mode Decomposition.

Many physical processes display complex high-dimensional time-varying behavior, from global weather patterns to brain activity. An outstanding challenge is to express high dimensional data in terms of a dynamical model that reveals their spatiotemporal structure. Dynamic Mode Decomposition is a means to achieve this goal, allowing the identification of key spatiotemporal modes through the diagonalization of a finite dimensional approximation of the Koopman operator. However, these methods apply best to time-translationally invariant or stationary data, while in many typical cases, dynamics vary across time and conditions. To capture this temporal evolution, we developed a method, Non-Stationary Dynamic Mode Decomposition, that generalizes Dynamic Mode Decomposition by fitting global modulations of drifting spatiotemporal modes. This method accurately predicts the temporal evolution of modes in simulations and recovers previously known results from simpler methods. To demonstrate its properties, the method is applied to multi-channel recordings from an awake behaving non-human primate performing a cognitive task.

Distributed Regulation of the Safety Factor Profile in Tokamaks Using Nonlinear Infinite-dimensional Control

Tokamaks are toroidal devices that use helical magnetic fields to confine a plasma (hot ionized gas). Such confinement increases the probability of ionic collisions. When the colliding ions have high enough kinetic energy, they can overcome the Coulombic forces of repulsion and fuse to form a heavier ion. The difference in mass between reactant and product ions is turned into energy, which can potentially be harvested to meet the growing world's energy demands. In tokamaks, the safety factor profile is a plasma property that characterizes the pitch of the helical magnetic field. Experiments have shown that the safety factor is related to the magnetohydrodynamic (MHD) stability of the confined plasma as well as to the capability of achieving highly-confined steady-state operation. Thus, active control of the safety factor profile or related plasma properties is critical for achieving MHD-stable, high-performance plasma operation. The evolution of the safety factor profile is governed by a nonlinear nonautonomous partial differential equation (PDE). The most commonly used control design approach reduces the governing PDE to a set of ordinary differential equations (ODEs) before synthesizing a control law. In this work, a distributed nonlinear safety-factor control law that does not require a finite-dimensional approximation of the governing PDE is proposed. Rigorous analysis shows that the proposed control law can drive the error between the safety factor profile and target profile to zero. The effectiveness of the proposed control law is demonstrated using nonlinear simulations for the DIII-D tokamak.

A Remark on Tonelli’s Calculus of Variations

This paper provides a quite simple method of Tonelli’s calculus of variations with positive definite and superlinear Lagrangians. The result complements the classical literature of calculus of variations before Tonelli’s modern approach. Inspired by Euler’s spirit, the proposed method employs finite-dimensional approximation of the exact action functional, whose minimizer is easily found as a solution of Euler’s discretization of the exact Euler – Lagrange equation. The Euler – Cauchy polygonal line generated by the approximate minimizer converges to an exact smooth minimizing curve. This framework yields an elementary proof of the existence and regularity of minimizers within the family of smooth curves and hence, with a minor additional step, within the family of Lipschitz curves, without using modern functional analysis on absolutely continuous curves and lower semicontinuity of action functionals.

Model-based Estimator Design for the Curing Process of a Concrete Structure

Concrete is one of the most important building materials and is used as a collective term for different dispersions of cement and aggregates. High performance concrete (HPC) is a rather new type of concrete mixture. A topic of current research is the behavior of fresh concrete during heat treatment. It is known that the heat treatment of fresh concrete accelerates the reaction of the cement and thus the chemically complex process of hydration. The mathematical representation of the curing concrete is a further step towards the systematic investigation. The objective of this work is the development of a state estimator to determine the temperature behavior of the fresh HPC mixture under heat treatment. Using a continuum representation describing the spatial-temporal evolution of temperature, moisture, and maturity of fresh HPC during heat treatment a finite-dimensional approximation is derived using a high order finite difference scheme. Experimental data is included for model parameterization by optimization. Based on this, three different nonlinear extensions of the Kalman filter (KF) techniques are realized and compared combining simulated and experimental data.

FINITE APPROXIMATIONS OF THE PROBLEM OF OPTIMAL IMPULSE STABILIZATION FOR A SYSTEM WITH AFTEREFFECT

A method for constructing finite-dimensional approximations for the optimal impulse stabilizing control of a linear autonomous system of differential equations with aftereffect is proposed. Auxiliary finite-dimensional problems of optimal stabilization of linear autonomous systems of ordinary differential equations with non-impulse controls are used. A procedure for reducing the dimension of a system of algebraic matrix Riccati equations is described.

Mean-square Approximation of Iterated Ito and Stratonovich Stochastic Integrals: Method of Generalized Multiple Fourier Series. Application to Numerical Integration of Ito Sdes and Semilinear Spdes (third Edition)

This is the third edition of the monograph (first edition 2020, second edition 2021) devoted to the problem of mean-square approximation of iterated Ito and Stratonovich stochastic integrals with respect to components of the multidimensional Wiener process. The mentioned problem is considered in the book as applied to the numerical integration of non-commutative Ito stochastic differential equations and semilinear stochastic partial differential equations with nonlinear non-commutative trace class noise. The book opens up a new direction in researching of iterated stochastic integrals. For the first time we use the generalized multiple Fourier series converging in the sense of norm in Hilbert space for the expansion of iterated Ito stochastic integrals of arbitrary multiplicity k with respect to components of the multidimensional Wiener process (Chapter 1). Sections 1.11-1.13 (Chapter 1) are new and generalize the results of Chapter 1 obtained earlier by the author and are also closely related to the multiple Wiener stochastic integral introduced by Ito in 1951. The convergence with probability 1 as well as the convergence in the sense of n-th (n=2, 3,...) moment for the expansion of iterated Ito stochastic integrals have been proved (Chapter 1). Moreover, the rate of both types of convergence has been established. The main difference between the third and second editions of the book is that the third edition includes original material (Chapter 2, Sections 2.10-2.19) on a new approach to the series expansion of iterated Stratonovich stochastic integrals of arbitrary multiplicity k with respect to components of the multidimensional Wiener process. The above approach allowed us to generalize some of the author's earlier results and also to make significant progress in solving the problem of series expansion of iterated Stratonovich stochastic integrals. In particular, for iterated Stratonovich stochastic integrals of the fifth and sixth multiplicity, series expansions based on multiple Fourier-Legendre series and multiple trigonometric Fourier series are obtained. In addition, expansions of iterated Stratonovich stochastic integrals of multiplicities 2 to 4 were generalized. These results (Chapter 2) adapt the results of Chapter 1 for iterated Stratonovich stochastic integrals. Two theorems on expansion of iterated Stratonovich stochastic integrals of arbitrary multiplicity k based on generalized iterated Fourier series with pointwise convergence are formulated and proved (Chapter 2). The results of Chapters 1 and 2 can be considered from the point of view of the Wong-Zakai approximation for the case of a multidimensional Wiener process and the Wiener process approximation based on its series expansion using Legendre polynomials and trigonometric functions. The integration order replacement technique for iterated Ito stochastic integrals has been introduced (Chapter 3). Exact expressions are obtained for the mean-square approximation error of iterated Ito stochastic integrals of arbitrary multiplicity k (Chapter 1) and iterated Stratonovich stochastic integrals of multiplicities 1 to 4 (Chapter 5). Furthermore, we provided a significant practical material (Chapter 5) devoted to the expansions and mean-square approximations of specific iterated Ito and Stratonovich stochastic integrals of multiplicities 1 to 6 from the unified Taylor-Ito and Taylor-Stratonovich expansions (Chapter 4). These approximations were obtained using Legendre polynomials and trigonometric functions. The methods constructed in the book have been compared with some existing methods (Chapter 6). The results of Chapter 1 were applied (Chapter 7) to the approximation of iterated stochastic integrals with respect to the finite-dimensional approximation of the Q-Wiener process (for integrals of multiplicity k) and with respect to the infinite-dimensional Q-Wiener process (for integrals of multiplicities 1 to 3).

Deep Neural Networks With Koopman Operators for Modeling and Control of Autonomous Vehicles

Autonomous driving technologies have received notable attention in the past decades. In autonomous driving systems, identifying a precise dynamical model for motion control is nontrivial due to the strong nonlinearity and uncertainty in vehicle dynamics. Recent efforts have resorted to machine learning techniques for building vehicle dynamical models, but the generalization ability and interpretability of existing methods still need to be improved. In this paper, we propose a pure data-driven vehicle modeling approach based on deep neural networks with an interpretable Koopman operator. The main advantage of using the Koopman operator is to represent the nonlinear dynamics in a linear lifted feature space. In the proposed approach, a deep learning-based extended dynamic mode decomposition algorithm is presented to learn a finite-dimensional approximation of the Koopman operator. A multi-step prediction loss function is used in the training process, enabling a long-term prediction capability. Furthermore, a data-driven model predictive controller with the learned Koopman model is designed for velocity profile tracking control of autonomous vehicles. Simulation results in a high-fidelity CarSim environment show that our approach outperforms previously developed traditional and advanced modeling methods. Velocity profile tracking tests of the autonomous vehicle are also performed in the CarSim environment. The results show that our approach has better tracking accuracy and higher computational efficiency than the model predictive control algorithms using a nonlinear model and a linear time-varying model.