An interior point algorithm with inexact step computation in function space for state constrained optimal control (Q644776)

This paper is concerned with optimal control problems governed by partial differential equations, subject to pointwise state constraints. An interior point method in function space is considered. The construction and analysis of an algorithm that integrates a Newton path-following method with adaptive grid refinement are made. A general result for the representation of the error in a suitable norm in function space is demonstrated. This leads to the definition of an a-posteriori error estimator for the error using the norm above. The overall construction of the algorithm is very well explained and analysed. Interesting numerical experiments are presented and analysed.