Forward-looking variables in deterministic control (Q2365109)

A classical LQ (linear object, quadratic criterion) discrete-time optimization procedure is iteratively applied to find best values of specific deterministic forward-looking variables of system equations. In the first iteration the standard LQ problem is solved. From the beginning of the second iteration, the equation for the controlled object is different -- it contains exogenous variables with values taken from the previous optimal trajectory iteration (nominal path of exogenous variables). The chosen simulation examples demonstrate the applicability of the proposed method to a broad class of macroeconomic control problems.