Error estimates for the numerical approximation of Neumann control problems governed by a class of quasilinear elliptic equations (Q694519)

The convergence behavior of finite element approximations of Neumann boundary optimal control problems governed by quasilinear equations is studied. Because the second order derivative operator depends on the state variable, the state equation is not monotone. Difficulties arise from the non-monotone character of the state equation. Existence and uniqueness of the discrete solution are discussed. The controls are discretized by piecewise constant functions. The estimated order of convergence in the \(L^2\)-norm for problems in convex domains is 1 and this is confirmed numerically. The numerically obtained convergence rates for a problem in an L-shaped (non-convex) domain is higher than the predicted one \({\mathcal O} (h^{1/2})\). The paper makes an important contribution to the analysis of the behavior of discrete solutions of this kind of optimal control problems which are more difficult than the distributed ones and with monotone state equations.