Abstract
In this paper, a parabolic boundary control problem with linear state equation and quadratic coercive objective functional is considered. It is shown, how the optimal control depends on the regularization parameter ν < 0, which is responsible for the coercivity of the objective functional. A new notion of switching points is introduced and their convergence for ν → + 0 is investigated. Numerical examples illustrate the theory.
AMS 1980 Subject Classifications: