Approaches to the optimizing control problem

The selection of the steady-state controls which enable a system to operate in an optimum manner is the optimizing control problem. An examination of direct and adaptive model-based approaches to this problem is made. In the direct approach, system measurements are used directly within an optimization procedure while, with the adaptive model-based techniques, an updated model is used for the optimization problem and the results of the model-based problem are applied to the system. Here, due to its recognized efficiency among established optimization algorithms, a conjugate direction method is adopted for the direct approach. For the model-based approach, two different techniques are described. A rationalized form of modified two-step method is presented along with a new method based on an approximate linear model. An assessment of the efficiency of the various methods is made through the simulation of a gas mixer system. The direct method is seen to be unrealistic while, for the system considered, the approximate linear model approach is seen to be superior to the modified two-step method