Optimal constrained restoration and dynamic programming

Recently, we have proposed a deconvolution procedure based on the optimal control theory. Two approaches have been made : first in the unconstrained case an analytical solution have been given, second, in order to include some constraints on the input signal to be restored, we have developed an iterative procedure. In fact, these two methods proceed of the application of the calculus of variation. In this paper we propose to use an other point of view to perform optimal restoration with constraints on the input signal and/or the state vector.