A study of weighting factors of the quadratic performance index

Date

1968

Journal Title

Journal ISSN

Volume Title

Publisher

Abstract

Although optimization procedures are attractive from the theoretical viewpoint, sometimes the resulting design has unsatisfactory performance characteristics. The use of weighting factors in the generalized quadratic performance index gives the designer the possibility of guiding the result. Yet, there has been published very little, if any, guidance in the selection of weighting factors. As a result of empirical research, it is shown that weighting factors are directly related to the slopes of the elements of the cost functional. They are defined in terms of these slopes. Uniqueness and non-uniqueness are discussed and related to the number of weighting factors and the order of optimization. The general conditions for the extremal which is determined by chosen weighting factors to be minimal are presented. In some cases, these conditions are expressed as a function of the system parameters and weighting factors. The concept of negative weighting factors is presented. In some examples, the usefulness of negative weighting factors is illustrated. General relationships between weighting factors and system parameters are developed analytically. These relationships are applied to second, third and fourth order systems. The application of these general relationships results in specific graphs and tables as presentations of the relationships between weighting factors and system parameters and weighting factors and time response performance measures. Convenient formulations for the generalized indices ISE and ITSE are derived and utilized in an empirical analysis of a broad class of systems. Several successful rules for weighting factors are developed with the objective of minimizing the maximum overshoot resulting from cascade lead network compensation.

Description

Keywords

Citation