A mixed finite element scheme for optimal control problems with pointwise state constraints (Q639398): Difference between revisions

The authors consider the optimal control problems with point-wise state constraints. Many numerical strategies to solve this problem are mentioned in the introduction. They are all based on the first order optimality conditions. The proposed paper uses another approach whose main idea consists in substituting the control variable in the objective according to the equation constraint. The reformulated problem is a constrained minimization problem involving only state variable and is equivalent to a fourth order variational inequality. The equivalence of all formulations can be established using the results of \textit{R. Glowinski, J. L. Lions} and \textit{R. Tremolieres} [Numerical analysis of variational inequalities. Studies in Mathematics and its Applications, Vol. 8. Amsterdam, New York, Oxford: North-Holland Publishing Company (1981; Zbl 0463.65046)] and of \textit{W. Liu, W. Gong} and \textit{N. N. Yan} [J. Comput. Math. 27, No. 1, 97--114 (2009; Zbl 1199.49067)]. To solve the reformulated problem, the authors adopt a new approach, the mixed variational scheme and its finite element approximation. The weak form, the mixed variational form and the discrete form of the problem where the control and the state are discretized by piecewise linear functions are derived. The finite element approximation is reformulated as the variational inequality problem and this problem is solved using a projection gradient method introduced by \textit{B. S. He} [Numer. Math. 68, No. 1, 71--80 (1994; Zbl 0822.65040)]. The main part of the paper is devoted to a priori and a posteriori error analysis of the mixed finite element scheme.