A discrete optimal control problem for descriptor systems

A class of discrete optimization problems described by nonlinear descriptor systems is considered on the basis of nonlinear programming. A discrete maximum principle is derived for a very general structural form of descriptor systems. The necessary and sufficient conditions for optimality are given for a subclass of optimization problems which is represented by linear descriptor dynamics with quasiconvex inequality constraints and pseudoconvex cost functional. For illustration, an example of a county macroeconomic system is presented.<
>