Synergy of predictive control and identification

The paper briefly explains how synergy works for minimum variance (MV) and pole placement control (PPC). A general definition of synergy between control and identification is given. Then synergy is studied for model predictive control (MPC), both in simulations and in theory. The investigation for MPC is made more difficult by that the previous results from MV and PPC do not carry over as the controller transfer function does not depend on the model transfer function point by point on the frequency scale but the controller rather depends on the whole model estimated. Finally, the role of dual predictive control is discussed relative to synergistic interaction of closed-loop identification and control