Mesh-independence of semismooth Newton methods for Lavrentiev-regularized state constrained nonlinear optimal control problems (Q2480886)

The authors consider the following mixed control-state constrained, or equivalently Lavrentiev-regularized state constrained, optimal control problem \[ \text{minimize }J(u,y):= \tfrac{1}{2}\| y- y_d\|^2_{L^2}+ \tfrac{\alpha}{2}\| u\|^2_{L^2} \] over \((u,y)\in L^2(\Omega)\times H^1_0(\Omega)\cap C(\Omega)\) \[ \text{subject to }Ay+ d(\cdot, y)= u\quad\text{in }\Omega, \] where \(A\) is a second-order linear elliptic differential operator and \[ y_a\leq \varepsilon u(x)+ y(x)\leq y_b\quad\text{for almost all }x\in\Omega. \] Based on the first-order necessary optimality conditions for this problem, a semi-smooth Newton method is proposed and ist fast local convergence in function space as well as a mesh-independence principle for appropriate discretizations are proved. Numerical tests are presented.