scientific article; zbMATH DE number 1006854
The paper is a survey of results by the authors related to numerical approximations of minimax (or/and viscosity) solutions in the Cauchy problem for the Hamilton-Jacobi equations. \[ \partial w(t,x)/\partial t + H(t,x,D_{x} w(t,x))=0, \;w(\theta, x) = \sigma(x),\quad (t,x)\in (0,\theta)\times R^n\tag{1} \] where the Hamiltonian is \[ H(t,x,p)=\max_{v\in Q} \langle f^{1} (t,x,v), p \rangle + \min_{u\in P} \langle f^{2} (t,x,u),p \rangle.\tag{2} \] The symbol \(D_{x} w\) denotes the gradient of the function \(w\), the symbol \(\langle f,p \rangle\) denotes the inner product of vectors \(f\) and \(p\), sets \(P\subset R^p\), \(P\subset R^q\) are compacta. The considered numerical approximations are based on the definition of the minimax solution as a continuous function \((0,\theta] \times R^n \ni (t,x) \mapsto u(t,x) \in R\), whose epigraph (hypograph) is weakly invariant relative to the family of upper (lower) characteristic inclusions. The property of weak invariance is provided in the constructions with the help of stable absorbtion operators (SAO) which are widely applied in differential games theory. One-step approximations of SAO are introduced at step \(\Delta\) on the time interval partition relative to the local convex hull of the epigraph of the continuous function \(\xi(x): R^n \to R\), or to the local concave hull of the hypograph of \(\xi(x)\), or to local linear approximation for the graph of a grid function \(\xi(x_{i})\) constructed by the method of least squares. Formulae for the one-step approximations of \(\xi(t,\cdot)\) are utilizing the sub- or super-differentials of \(\xi(t+\Delta ,y)\) in a small neighborhood of \(x\). The operators are the base of step-by-step backward procedures or the so-called operators of sequential maximum (OSM). It is proved that results of OSM converge to the minimax solutions with the rate \(\Delta^{1/2}\). Approximation schemes on the base of these operators are obtained for grid functions. There is a comparison of the approximations and known schemes. Illustrating examples of numerical calculations are given.
- Nonsmooth analysis and Hamilton--Jacobi equations on Riemannian manifolds
- Analytic-numerical approach to construction of minimax solution to the Hamilton-Jacobi equation in three-dimensional space
- Numerical approximation of the maximal solutions for a class of degenerate Hamilton-Jacobi equations
- scientific article; zbMATH DE number 1023200 (Why is no real title available?)
- The use of nonsmooth analysis and of duality methods for the study of Hamilton-Jacobi equations
This page was built for publication:
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4336015)