Computational and performance aspects of static decentralized controllers

This paper presents a computational approach to the problem of determining the optimal constant feedback gain matrices for a particular decentralized controller structure. The technique presented utilizes an extension of the concept of ε-coupling, which has been used successfully for large-scale centralized structures. The results reported herein represent a refinement of the authors' previous work in developing decentralized controllers, into an algorithmic format. The effectiveness of this technique is discussed, and guidelines for its use are given. The algorithm is also applied to determine the optimal feedback gains for a two-area power system, and the performance of the resulting controller is compared to that of other types of decentralized controllers having the same structure.