Multivariable Self-Tuning Controllers Applicable to Systems with Measurements of Differing Time Delays

Discrete multiple-input multiple-output linear systems with measurements of differing time delays are considered. By suitable manipulations, the original system model is rewritten in three alternative forms, all of which lead to a direct calculation of the optimal control policy. By combining a recursive least squares parameter estimator with the estimated model optimal control law for each model form, three alternative self-tuning control schemes are obtained. Each one uses a different amount of the most recently obtained information and involves different computational requirements. Which one is preferable depends on the number of inputs, number of outputs, system order and time delays of the process to be controlled, as well as on the capabilities of the available computer. All three self-tuning control algorithms are shown to be successful in controlling a continuous-flow stirred tank reactor model.