As we know, the theory about the linear equation AX−XB=C has already been well developed in the finite-dimensional cases. In this paper, we will try to extend it to infinite-dimensional cases by using a similar technique developed recently in the finite-dimensional case.

In this dissertation we study the Lyapunov diagonal stability and its extensions through partitions of the index set {1,...,n}. This type of matrix stability plays an important role in various applied areas such as population dynamics, systems theory and complex networks. We first examine a result of Redheffer that reduces Lyapunov diagonal stability of a matrix to common diagonal Lyapunov solutions on two matrices of order one less. An enhanced statement of this result based on the Schur complement formulation is presented here along with a shorter and purely matrix-theoretic proof. We develop a number of extensions to this result, and formulate the range of feasible common diagonal Lyapunov solutions. In particular, we derive explicit algebraic conditions for a set of 2 x 2 matrices to share a common diagonal Lyapunov solution. In addition, we provide an affirmative answer to an open problem concerning two different necessary and sufficient conditions, due to Oleng, Narendra, and Shorten, for a pair of 2 x 2 matrices to share a common diagonal Lyapunov solution. Furthermore, the connection between Lyapunov diagonal stability and the P-matrix property under certain Hadamard multiplication is extended. Accordingly, we present a new characterization involving Hadamard multiplications for simultaneous Lyapunov diagonal stability on a set of matrices. In particular, the common diagonal Lyapunov solution problem is reduced to a more convenient determinantal condition. This development is based upon a new concept called P-sets and a recent result regarding simultaneous Lyapunov diagonal stability by Berman, Goldberg, and Shorten. Next, we consider various types of matrix stability involving a partition alpha of {1,..., n}. We introduce the notions of additive H(alpha)-stability and P_0(alpha)-matrices, extending those of additive D-stability and nonsingular P_0-matrices. Several new results are developed, connecting additive H(alpha)-stability and the P_0(alpha)-matrix property to the existing results on matrix stability involving alpha. We also point out some differences between these types of matrix stability and the conventional matrix stability. Besides, the extensions of results related to Lyapunov diagonal stability, D-stability, and additive D-stability are discussed. Finally, we introduce the notion of common alpha-scalar diagonal Lyapunov solutions over a set of matrices, which is a generalization of common diagonal Lyapunov solutions. We present two different characterizations of this new concept based on the well-known results for Lyapunov alpha-scalar stability [34]. The first one hinges on a general version of the theorem of the alternative, and the second one using Hadamard multiplications stems from an extension of the P-set property. Several illustrative examples and an application concerning a set of block upper triangular matrices are provided.

We study numerical methods for finding the maximal symmetric positive definite solution of the nonlinear matrix equation $X = Q + LX^{-1}L^T$, where Q is symmetric positive definite and L is nonsingular. Such equations arise for instance in the analysis of stationary Gaussian reciprocal processes over a finite interval. Its unique largest positive definite solution coincides with the unique positive definite solution of a related discrete-time algebraic Riccati equation (DARE). We discuss how to use the butterfly SZ algorithm to solve the DARE. This approach is compared to several fixed point type iterative methods suggested in the literature.

We study numerical methods for finding the maximal symmetric positive definite solution of the nonlinear matrix equation $X = Q + LX^{-1}L^T$, where Q is symmetric positive definite and L is nonsingular. Such equations arise for instance in the analysis of stationary Gaussian reciprocal processes over a finite interval. Its unique largest positive definite solution coincides with the unique positive definite solution of a related discrete-time algebraic Riccati equation (DARE). We discuss how to use the butterfly SZ algorithm to solve the DARE. This approach is compared to several fixed point type iterative methods suggested in the literature.

We present a new method for the computation of low rank approximations to the solution of large, sparse, stable Lyapunov equations. It is based on a generalization of the classical Smith method and profits by the usual low rank property of the right hand side matrix. The requirements of the method are moderate with respect to both computational cost and memory. Hence, it provides a possibility to tackle large scale control problems. Besides the efficient solution of the matrix equation itself, a thorough integration of the method into several control algorithms can improve their performance to a high degree. This is demonstrated for algorithms for model reduction and optimal control. Furthermore, we propose a heuristic for determining a set of suboptimal ADI shift parameters. This heuristic, which is based on a pair of Arnoldi processes, does not require any a priori knowledge on the spectrum of the coefficient matrix of the Lyapunov equation. Numerical experiments show the efficiency of the iterative scheme combined with the heuristic for the ADI parameters.

We present a new method for the computation of low rank approximations to the solution of large, sparse, stable Lyapunov equations. It is based on a generalization of the classical Smith method and profits by the usual low rank property of the right hand side matrix. The requirements of the method are moderate with respect to both computational cost and memory. Hence, it provides a possibility to tackle large scale control problems. Besides the efficient solution of the matrix equation itself, a thorough integration of the method into several control algorithms can improve their performance to a high degree. This is demonstrated for algorithms for model reduction and optimal control. Furthermore, we propose a heuristic for determining a set of suboptimal ADI shift parameters. This heuristic, which is based on a pair of Arnoldi processes, does not require any a priori knowledge on the spectrum of the coefficient matrix of the Lyapunov equation. Numerical experiments show the efficiency of the iterative scheme combined with the heuristic for the ADI parameters.