Sparse solution of the linear quadratic regulator problem for large scale systems applications

Abstract

This paper presents a computationally efficient method for calculating the optimal feedback gain applicable to large scale systems. The plant is represented by using descriptor systems and the gain is calculated directly through the numerical integration of the Chandrasekhar equations. This allows the application of modern sparsity techniques, thus reducing the computational burden. Tests in two dynamic systems are conducted in order to assess the performance of the proposed method as compared with other techniques based on the conventional solution of the algebraic Riccati equation.