A Neural Network-Based Policy Iteration Algorithm with Global $$H^2$$ -Superlinear Convergence for Stochastic Games on Domains

In this work, we propose a class of numerical schemes for solving semilinear Hamilton–Jacobi–Bellman–Isaacs (HJBI) boundary value problems which arise naturally from exit time problems of diffusion processes with controlled drift. We exploit policy iteration to reduce the semilinear problem into a sequence of linear Dirichlet problems, which are subsequently approximated by a multilayer feedforward neural network ansatz. We establish that the numerical solutions converge globally in the \(H^2\)-norm and further demonstrate that this convergence is superlinear, by interpreting the algorithm as an inexact Newton iteration for the HJBI equation. Moreover, we construct the optimal feedback controls from the numerical value functions and deduce convergence. The numerical schemes and convergence results are then extended to oblique derivative boundary conditions. Numerical experiments on the stochastic Zermelo navigation problem are presented to illustrate the theoretical results and to demonstrate the effectiveness of the method.

     

Understanding the solubility and crystallization of molybdenum in high-sodium borosilicate glasses: effect of lanthanum oxide

Journal of Radioanalytical and Nuclear Chemistry - The simplified 8-oxide system (SiO2-B2O3-Na2O-Al2O3-CaO-TiO2-MoO3-La2O3) has been prepared under two types of cooling methods, and analyzed with...     

Anti-fibrosis attributes: UHPLC-MS/MS-based pharmacokinetics profiling of a novel autotaxin inhibitor with excellent in vivo efficacy in rat

3,4-Difluorobenzyl(1-ethyl-5-(4-((4-hydroxypiperidin-1-yl)-methyl)thiazol-2-yl)-1H-indol-3-yl)carbamate (NAI59), a small molecule with outstanding therapeutic effectiveness to anti-pulmonary fibrosis, was developed as an autotaxin inhibitor candidate compound. To evaluate the pharmacokinetics and plasma protein binding of NAI59, a UPLC–MS/MS method was developed to quantify NAI59 in plasma and phosphate-buffered saline. The calibration curve linearity ranged from 9.95 to 1990.00 ng/mL in plasma. The accuracy was −6.8 to 5.9%, and the intra- and inter-day precision was within 15%. The matrix effect and recovery, as well as dilution integrity, were within the criteria. The chromatographic and mass spectrometric conditions were also feasible to determine phosphate-buffered saline samples, and it has been proved that this method exhibits good precision and accuracy in the range of 9.95–497.50 ng/mL in phosphate-buffered saline. This study is the first to determine the pharmacokinetics, absolute bioavailability, and plasma protein binding of NAI59 in rats using this established method. Therefore, the pharmacokinetic profiles of NAI59 showed a dose-dependent relationship after oral administration, and the absolute bioavailability in rats was 6.3%. In addition, the results of protein binding showed that the combining capacity of NAI59 with plasma protein attained 90% and increased with the increase in drug concentration.     

CCSD(T) study on the structures and chemical bonds of AnO molecules (An = Bk–Lr)

The molecular geometries and dissociation energies of AnO (An = Bk–Lr) molecules were first obtained at thecoupled-cluster single-, double-, and perturbative triple-excitations [CCSD(T)] level of theory. Four hybrid functionals,B3LYP, M06-2X, TPSSh, and PBE0, were also employed in the calculations for the sake of comparison. In comparison ofthe CCSD(T) results, B3LYP, TPSSh, and PBE0 functionals can obtain more appropriate results than M06-2X and MP2.The analyses on molecular orbitals show that the 7s, 6d, and 5f atomic orbitals of actinide (An) atoms participate in thebonding of An–O bonds. The partial covalent nature between An and O atoms is revealed by QTAIM analyses.     

Optimal parameter selection for constraint-following control for mechanical systems based on Stackelberg game

Nonlinear Dynamics - This paper proposes an optimal parameter design of control scheme for mechanical systems by adopting the Stackelberg game theory. The goal of the control is to drive the...     

Model and nonlinear dynamic analysis of linear guideway subjected to external periodic excitation in five directions

Nonlinear Dynamics - Investigation of the dynamic characteristics of linear guideway is essential for enhancing structural performance of machine tools. However, few mechanical models are...     

Majorana Representation for a Composite System

International Journal of Theoretical Physics - The Majorana representation, which provides an intuitive way to represent the quantum state by stars on the Bloch sphere, has drawn considerable...     

Restricted testing conditions for the multilinear maximal operator

Restricted testing conditions were considered recently. For the maximal operator, Hytönen, Li and Sawyer [8] first obtained parental testing condition. Later, they [9] showed that it suffices to restrict testing to doubling cubes. Chen and Lacey [3] gave a similar restricted testing condition. In our paper, we discuss a version of the latter in the multilinear setting.

     

From differential equation solvers to accelerated first-order methods for convex optimization

Mathematical Programming - Convergence analysis of accelerated first-order methods for convex optimization problems are developed from the point of view of ordinary differential equation solvers. A...