Linear controller design by goal programming

A computational optimization-based approach for designing fixed-order controllers is presented. The approach focuses classical pole placement design techniques for both certain and uncertain (interval) linear time-invariant plants within a goal programming framework. The design objective is the minimization of the overall deviation from the desired performance for the closed-loop system, as specified by a polytope of characteristic polynomials in the general case. Goal programming, a traditional technique in many different engineering design contexts, reveals to be an adequate basis for integrating a number of existing results in linear controller design by pole placement. Examples illustrate the main characteristics of the proposed approach.