In this paper, we present a novel penalty approach for the numerical solution of continuously controlled HJB equations and HJB obstacle problems. Our results include estimates of the penalization error for a class of penalty terms, and we show that variations of Newton’s method can be used to obtain globally convergent iterative solvers for the penalized equations. Furthermore, we discuss under...

An approximately globally convergent numerical method for a 1-D Coefficient Inverse Problem for a hyperbolic PDE is applied to image dielectric constants of targets from blind experimental data. The data were collected in the field by the Forward Looking Radar of the US Army Research Laboratory. A posteriori analysis has revealed that computed and tabulated values of dielectric constants are in...

We consider the local behaviour of inexact Newton methods for the solution of a semis-mooth system of equations. In particular, we give a complete characterization of the Q-superlinear and Q-quadratic convergence of inexact Newton methods. We then apply these results to a particular semismooth system of equations arising from variational inequality problems, and present a globally and locally f...

In this paper, we proposed a modified Logarithmic-Quadratic Proximal (LQP) method [Auslender et al.: Comput. Optim. Appl. 12, 31–40 (1999)] for solving variational inequalities problems. We solved the problem approximately, with constructive accuracy criterion. We show that the method is globally convergence under that the operator is pseudomonotone which is weaker than the monotonicity and the...

We show that two popular discounted reward natural actor-critics, NAC-LSTD and eNAC, follow biased estimates of the natural policy gradient. We derive the first unbiased discounted reward natural actor-critics using batch and iterative approaches to gradient estimation and prove their convergence to globally optimal policies for discrete problems and locally optimal policies for continuous prob...

In this paper, we combine the inexact Newton method with the stochastic decomposition method and present a stochastic Newton method for solving two-stage stochastic programs. We prove that the new method is globally convergent and, if in addition, an integral approximation error bound condition holds, the convergence is superlinear (with probability one for random integration rules). The error ...

We consider a multidimensional model for the dynamic combustion of compressible reacting fluids formulated by the Navier Stokes equations in Euler coordinates. For the chemical model we consider a one way irreversible chemical reaction governed by the Arrhenius kinetics. The existence of globally defined weak solutions of the Navier-Stokes equations for compressible reacting fluids is establish...

In this paper, we derive an unconstrained convex programming approach to solving standard form linear programs through an entropic perturbation. The whole duality theory is established by using only one simple inequality "In z S; z 1 for z > 0". A curved search algorithm is also proposed for obtaining a pair of primal and dual G-optimal solutions. The proposed algorithm is proven to be globally...

This note deals with chained form systems with strongly nonlinear unmodeled dynamics and external disturbances. The objective is to design a robust nonlinear state feedback law such that the closed-loop system is globally -exponentially stable. We propose a novel switching control strategy involving the use of input/state scaling and integrator backstepping. The new features of our controllers ...

Conventional Blind Signal Separation algorithms do not adopt any asymmetric information of the input sources, thus the convergence point of a single output is always unpredictable. In this paper, a new Principal Independent Component Analysis concept is proposed, we try to extract the objective Independent Component directly without separating all the signals. A cumulant-based globally converge...