Differentiability of the value function in continuous-time economic models (Q439265)

The authors consider an infinite-horizon optimization problem in continuous time, which has a lot of applications in economics. Their aim is, to give conditions for the differentiability of the value function and for the uniqueness of dual variables belonging to the superdifferential of the value function. Considering the existing literature up to this paper the point is, to dispense with the assumption of inner optimal paths. There is an earlier paper of both the authors [J. Econ. Theory 144, No. 5, 1948--1964 (2009; Zbl 1195.90092)], which already works without those interior paths, but in discrete time optimization, and there is another earlier paper by \textit{L. M. Benveniste} and \textit{J. A. Scheinkman} [Econometrica 47, 727--732 (1979; Zbl 0435.90031)], which uses continuous time, but with the inner point condition. So the authors, using infinite-dimensional calculus, extend the envelope theorem from the paper of Benveniste and Scheinkman to optimization problems with constraints and show, that more and stronger important results than in the model with discrete time are possible as for instance, that the differentiability of the value function at an initial point implies differentiability along the whole optimal trajectory. Results about Bellman's equation and examples from economics are added.