A structure-preserving doubling algorithm for solving a class of quadratic matrix equation with $M$-matrix

Cairong Chen
DOI: https://doi.org/10.3934/era.2022030
2021-03-12
Abstract:Consider the problem of finding the maximal nonpositive solvent $\Phi$ of the quadratic matrix equation (QME) $X^2 + BX + C =0$ with $B$ being a nonsingular $M$-matrix and $C$ an $M$-matrix such that $B^{-1}C\ge 0$, and $B - C - I$ a nonsingular $M$-matrix. Such QME arises from an overdamped vibrating system. Recently, Yu et al. ({\em Appl. Math. Comput.}, 218: 3303--3310, 2011) proved that $\rho(\Phi)\le 1$ for this QME. In this paper, we slightly improve their result and prove $\rho(\Phi)< 1$, which is important for the quadratic convergence of the structure-preserving doubling algorithm. Then, a new globally monotonically and quadratically convergent structure-preserving doubling algorithm to solve the QME is developed. Numerical examples are presented to demonstrate the feasibility and effectiveness of our method.
Numerical Analysis
What problem does this paper attempt to address?
The problem that this paper attempts to solve is to find the maximal nonpositive solvent of a specific class of quadratic matrix equations (QME). Specifically, the form of this quadratic matrix equation is: \[ X^2 + BX + C = 0 \] where \( B \) is a nonsingular M - matrix, \( C \) is an M - matrix, and satisfies \( B^{-1}C \geq 0 \), and \( B - C - I \) is also a nonsingular M - matrix. This type of quadratic matrix equation usually appears in overdamped vibration systems. ### Main contributions of the paper 1. **Improve existing results**: - Yu et al. (in 2011) proved that under the above conditions, the spectral radius \( \rho(\Phi) \leq 1 \) of the maximal nonpositive solution \( \Phi \). - This paper further improves this result and proves that \( \rho(\Phi) < 1 \). This improvement is crucial for the quadratic convergence of the structure - preserving doubling algorithm. 2. **Develop a new algorithm**: - Based on the above - improved results, the author proposes a new globally monotonic and quadratically convergent structure - preserving doubling algorithm to solve this quadratic matrix equation. 3. **Numerical verification**: - The effectiveness and feasibility of the new algorithm are verified through numerical experiments, showing its advantages over other methods. ### Mathematical formulas - Quadratic matrix equation: \[ X^2 + BX + C = 0 \] - Spectral radius condition: \[ \rho(\Phi) < 1 \] - Iterative formulas in the structure - preserving doubling algorithm (simplified representation): \[ E_{i + 1} = E_i (I_n - Y_i X_i)^{-1} E_i, \] \[ F_{i + 1} = F_i (I_n - X_i Y_i)^{-1} F_i, \] \[ X_{i + 1} = X_i + F_i (I_n - X_i Y_i)^{-1} X_i E_i, \] \[ Y_{i + 1} = Y_i + E_i (I_n - Y_i X_i)^{-1} Y_i F_i. \] ### Summary This paper solves the problem of finding the maximal nonpositive solution of a specific quadratic matrix equation by improving existing theoretical results and developing a new efficient algorithm, and verifies the effectiveness of the algorithm through numerical experiments.