EXTENSION OF HIERARCHICAL DYNAMIC CONTROL TO MULTIOBJECTIVE OPTIMIZATION.

The basic methodology of hierarchical dynamic control is extended to deal with multiobjective optimization problems. The class of problems considered is that of determining the optimal control strategy for linear interconnected dynamic systems with distinct agents against a finite performance criteria. A noninferior solution strategy is invoked to generate the Pareto-optimal controls. The computational procedure is implemented via a three-level hierarchical algorithm whereby the first level consisting of lower order, independent Riccati matrix and tracking vector equations, the second level specifying a coordinating mechanism to update the interaction patterns and at the third level, a steepest descent routine is used to generate the noninferior control strategies.