Model reduction for uncertain systems using LMIs

This paper presents a general criterion for solving reduction problem on uncertain systems by using LMIs. The criterion can be used for different uncertain systems, such as linear time-invarying systems, Markovian jump systems, and hybrid jump systems with continuous- and discrete-time cases so on. A novel loop algorithm is proposed to calculate reduced model. The algorithm convergent efficiently by making the use of binary search thought and parameter constraints. Some examples are given to illustrate the benefits of criterion and algorithm.