FINITE ELEMENT DISCRETIZATION OF STATE-CONSTRAINED ELLIPTIC OPTIMAL CONTROL PROBLEMS WITH SEMILINEAR STATE EQUATION

被引:33
|
作者
Neitzel, Ira [1 ]
Pfefferer, Johannes [2 ]
Roesch, Arnd [3 ]
机构
[1] Tech Univ Munich, Lehrstuhl M17, D-85748 Garching, Germany
[2] Univ Bundeswehr Munchen, Fak Bauingenieurwesen & Umweltwissensch, Inst Math & Bauinformat, D-85577 Neubiberg, Germany
[3] Univ Duisburg Essen, Fak Math, D-45127 Essen, Germany
基金
奥地利科学基金会;
关键词
optimal control; finite elements; semilinear elliptic PDE; state constraints; a priori error estimates; NUMERICAL APPROXIMATION; POINTWISE STATE; DIRICHLET PROBLEM; ERROR ANALYSIS; CONVERGENCE; SQP;
D O I
10.1137/140960645
中图分类号
TP [自动化技术、计算机技术];
学科分类号
0812 ;
摘要
We study a class of semilinear elliptic optimal control problems with pointwise state constraints. The purpose of this paper is twofold. First, we present convergence results for the finite element discretization of this problem class similarly to known results with finite-dimensional control space, thus extending results that are-for control functions-only available for linear-quadratic convex problems. We rely on a quadratic growth condition for the continuous problem that follows from second order sufficient conditions. Second, we show that the second order sufficient conditions for the continuous problem transfer to its discretized version. This is of interest, for example, when considering questions of local uniqueness of solutions or the convergence of solution algorithms such as the SQP method.
引用
收藏
页码:874 / 904
页数:31
相关论文
共 50 条