Abstract
This paper presents a review of the developments in the numerical solution of partial differential equations by linear programming since 1969. Since then, contributions have been essentially in four directions: (i) For linear problems, there has been rigorous theoretical investigation into the problem of convergence, as the number of basis functions tends to infinity and the minimization grid size tends to zero, and also into the problem of generating complete systems of basis functions for the general elliptic equation. (ii) For nonlinear problems, Newton’s method and Newton-type methods have been applied to quasilinear elliptic, parabolic and hyperbolic initial and boundary value problems with quasilinear boundary conditions. Convergence has been proved. Realistic error bounds and accurate computational methods have been derived. (iii) Interactive graphic display systems have been developed. (iv) $L_1 $ norm and irregular grids of discretization have been used.

This publication has 28 references indexed in Scilit: