Contribution to prior tuning of LQG selftuners

A prior prediction of control quality achievable by the optimally tuned LQG selftuner of a fixed structure is proposed. The prediction helps to judge in advance the usefulness of the intended selftuner implementation. The proposed algorithmization makes also possible the offline tuning of particular penalties in accordance with user's wishes. In this way, the used Bayesian methodology (together with existing solution of structure determination problem) provides algorithmic tools for systematic pre-tuning of majority of user's knobs