Abstract:We derive the alternating-directions implicit (ADI) method based on a commuting operator split and apply the results in detail to the continuous time algebraic Lyapunov equation with low-rank constant term and approximate solution, giving pointers for the Sylvester case. Previously, it has been mandatory to start the low-rank ADI for Lyapunov equations (CF-ADI, LR-ADI, G-LR-ADI) or Sylvester equations (fADI, G-fADI) with an all-zero initial value. Our approach extends the known efficient iteration schemes of low-rank increments and residuals to arbitrary low-rank initial values for all these methods. We further generalize two properties of the low-rank Lyapunov ADI to the generic ADI applied to arbitrary linear equations using a commuting operator split, namely the invariance of iterates under permutations of the shift parameters, and the efficient handling of complex shift parameters. We investigate the performance of arbitrary initial values using two outer iterations in which the low-rank Lyapunov ADI is typically called. First, we solve an algebraic Riccati equation with the Newton method. Second, we solve a differential Riccati equation with a first-order Rosenbrock method. Numerical experiments confirm that the proposed new initial value of the ADI can lead to a significant reduction in the total number of ADI steps, while also showing a 17% and 8x speed-up over the zero initial value for the two equation types, respectively.

What problem does this paper attempt to address?

The main problem that this paper attempts to solve is to improve the efficiency and flexibility of the Alternating Direction Implicit (ADI) method in solving linear matrix equations. Specifically, the author aims to extend the existing low - rank ADI methods so that they can handle non - zero initial values and are applicable to a wider range of linear equation types. ### Core Problems of the Paper 1. **Extending the ADI Method to Support Non - zero Initial Values**: - Traditionally, low - rank ADI methods (such as CF - ADI, LR - ADI, G - LR - ADI, etc.) must start from all - zero initial values when solving Lyapunov equations or Sylvester equations. This limits the flexibility and efficiency of the method. - The author proposes a new framework, enabling these methods to start from any low - rank initial values, thereby reducing the number of iterations and improving computational efficiency. 2. **Generalizing the Properties of the ADI Method**: - The author generalizes two important properties of low - rank Lyapunov ADI to more general linear equations: - The invariance of the iteration result to the permutation of shift parameters. - The existence of complex conjugate shift parameters and their corresponding real - valued two - step ADI methods. 3. **Improving Numerical Performance**: - By using improved initial values, the author shows that when solving the Algebraic Riccati Equation (ARE) and the Differential Riccati Equation (DRE), the new method can significantly reduce the total number of ADI steps and achieve a 17% and an 8 - fold speed - up respectively. ### Mathematical Formulas The key equations involved in the paper include: - **Continuous - time Algebraic Lyapunov Equation (ALE)**: \[ AX+XA^H = - GSG^H \] where \( A\in\mathbb{C}^{n\times n} \), \( G\in\mathbb{C}^{n\times g} \), \( S\in\mathbb{C}^{g\times g} \), and \( g\ll n \). - **Operator Splitting of the ADI Method**: \[ M_k(U)=(\alpha_k + \beta_k)^{-1}(A+\alpha_k I)U(A + \beta_k I)^H \] \[ N_k(U)=(\alpha_k+\beta_k)^{-1}(A-\beta_k I)U(A-\alpha_k I)^H \] - **Residual Update Formula**: \[ R_{k + 1}=R_k-2\text{Re}(\alpha_k)(A+\alpha_k I)^{-1}R_k \] ### Conclusion This paper successfully extends the ADI method to support non - zero initial values by introducing a fully commutative splitting scheme and applies it to a wider range of linear equation types. The experimental results show that this method not only improves computational efficiency but also provides new ideas and tools for solving complex linear equations.

A unifying framework for ADI-like methods for linear matrix equations and beneficial consequences

Low-rank generalized alternating direction implicit iteration method for solving matrix equations

Inexact linear solves in the low-rank ADI iteration for large Sylvester equations

A general alternating direction implicit iteration method for solving complex symmetric linear systems

A class of ADI method for the nonsymmetric coupled algebraic Riccati equations

Low-rank alternating direction doubling algorithm for solving large-scale continuous time algebraic Riccati equations

Low-Rank Methods for Solving Discrete-Time Projected Lyapunov Equations

Alternately linearized implicit iteration methods for the minimal nonnegative solutions of the nonsymmetric algebraic Riccati equations

Linearized Alternating Direction Method with Adaptive Penalty for Low-Rank Representation.

Linearized Alternating Direction Method with Adaptive Penalty for Low-Rank Representation

A Low-Rank ADMM Splitting Approach for Semidefinite Programming

Splitting ADI scheme for fractional Laplacian wave equations

Optimization methods for solving matrix equations

Preconditioned Low-Rank Riemannian Optimization for Symmetric Positive Definite Linear Matrix Equations

Robust Implicit Adaptive Low Rank Time-Stepping Methods for Matrix Differential Equations

Efficient second-order ADI difference schemes for three-dimensional Riesz space-fractional diffusion equations

Using $LDL^{T}$ factorizations in Newton's method for solving general large-scale algebraic Riccati equations

On improving the efficiency of ADER methods

An increasing rank Riemannian method for generalized Lyapunov equations