《《2016 Primal-dual interior point methods for PDE-constrained optimization》.pdf

PRIMAL-DUAL INTERIOR POINT METHODS FOR PDE-CONSTRAINED OPTIMIZATION MICHAEL ULBRICH∗ AND STEFAN ULBRICH† Abstract. This paper provides a detailed analysis of a primal-dual interior-point method for PDE-constrained optimiza- p tion. Considered are optimal control problems with control constraints in L . It is shown that the developed primal-dual interior-point method converges globally and locally superlinearly. Not only the easier L∞-setting is analyzed, but also a q ∞ more involved L -analysis, q ∞, is presented. In L , the set of feasible controls contains interior points and the Fr´echet q differentiability of the perturbed optimality system can be shown. In the L -setting, which is highly relevant for PDE- constrained optimization, these nice properties are no longer available. Nevertheless, a convergence analysis is developed using refined techniques. In particular, two-norm techniques and a smoothing step are required. 1. Introduction. This paper is concerned with the analysis of primal-dual interior point meth- ods for optimization problems with PDE- and pointwise inequality constraints. We assume that the problem has optimal control structure and that the inequality constraints are posed on the controls only. In contrast to state constraints, this situation allows for a rigorous analysis. Related investi- gations of other Newton-based algorithms were conducted in, e.g., [3, 7, 10, 11, 13, 17, 15, 16, 18, 19, 20] for comparable problem settings. For primal-dual interior point methods, although inten- sively investigated in finite dimensional mathematical programming, see, e.g., [6] and the