Table of Contents
Fetching ...

Convergence of two-stage iterative scheme for $K$-weak regular splittings of type II with application to Covid-19 pandemic model

Vaibhav Shekhar, Nachiketa Mishra, Debasisha Mishra

TL;DR

The paper develops convergence theory for a two-stage iterative scheme applied to $K$-weak regular splittings of type II within proper cone frameworks, extending classical results to nonstandard cones and inner-iteration splittings. It proves that, under a commutativity condition $VF^{-1}G=GF^{-1}V$, the stationary and non-stationary two-stage methods converge and the induced splitting remains $K$-weak regular type II, with the key matrices $T_{s(k)}$ and $\widehat{T}_{s(k)}$ being similar. It also derives comparison results showing how certain orderings of the splittings yield faster convergence, and establishes a monotone convergence theorem for $K$-nonnegative, $K$-monotone settings. The methods are then applied to a COVID-19 age-structured next-generation-matrix model to efficiently compute $R_0=\rho(\mathcal{B}\mathcal{A}^{-1})$, with numerical experiments demonstrating performance gains of the two-stage approach for both small and large-scale systems. Overall, the work provides a rigorous framework for fast, monotone convergence in cone-structured linear systems and practical tools for epidemiological risk assessment via next-generation matrices.

Abstract

Monotone matrices play a key role in the convergence theory of regular splittings and different types of weak regular splittings. If monotonicity fails, then it is difficult to guarantee the convergence of the above-mentioned classes of matrices. In such a case, $K$-monotonicity is sufficient for the convergence of $K$-regular and $K$-weak regular splittings, where $K$ is a proper cone in $\mathbb{R}^n$. However, the convergence theory of a two-stage iteration scheme in general proper cone setting is a gap in the literature. Especially, the same study for weak regular splittings of type II (even if in standard proper cone setting, i.e., $K=\mathbb{R}^n_+$), is open. To this end, we propose convergence theory of two-stage iterative scheme for $K$-weak regular splittings of both types in the proper cone setting. We provide some sufficient conditions which guarantee that the induced splitting from a two-stage iterative scheme is a $K$-regular splitting and then establish some comparison theorems. We also study $K$-monotone convergence theory of the stationary two-stage iterative method in case of a $K$-weak regular splitting of type II. The most interesting and important part of this work is on $M$-matrices appearing in the Covid-19 pandemic model. Finally, numerical computations are performed using the proposed technique to compute the next generation matrix involved in the pandemic model.

Convergence of two-stage iterative scheme for $K$-weak regular splittings of type II with application to Covid-19 pandemic model

TL;DR

The paper develops convergence theory for a two-stage iterative scheme applied to -weak regular splittings of type II within proper cone frameworks, extending classical results to nonstandard cones and inner-iteration splittings. It proves that, under a commutativity condition , the stationary and non-stationary two-stage methods converge and the induced splitting remains -weak regular type II, with the key matrices and being similar. It also derives comparison results showing how certain orderings of the splittings yield faster convergence, and establishes a monotone convergence theorem for -nonnegative, -monotone settings. The methods are then applied to a COVID-19 age-structured next-generation-matrix model to efficiently compute , with numerical experiments demonstrating performance gains of the two-stage approach for both small and large-scale systems. Overall, the work provides a rigorous framework for fast, monotone convergence in cone-structured linear systems and practical tools for epidemiological risk assessment via next-generation matrices.

Abstract

Monotone matrices play a key role in the convergence theory of regular splittings and different types of weak regular splittings. If monotonicity fails, then it is difficult to guarantee the convergence of the above-mentioned classes of matrices. In such a case, -monotonicity is sufficient for the convergence of -regular and -weak regular splittings, where is a proper cone in . However, the convergence theory of a two-stage iteration scheme in general proper cone setting is a gap in the literature. Especially, the same study for weak regular splittings of type II (even if in standard proper cone setting, i.e., ), is open. To this end, we propose convergence theory of two-stage iterative scheme for -weak regular splittings of both types in the proper cone setting. We provide some sufficient conditions which guarantee that the induced splitting from a two-stage iterative scheme is a -regular splitting and then establish some comparison theorems. We also study -monotone convergence theory of the stationary two-stage iterative method in case of a -weak regular splitting of type II. The most interesting and important part of this work is on -matrices appearing in the Covid-19 pandemic model. Finally, numerical computations are performed using the proposed technique to compute the next generation matrix involved in the pandemic model.

Paper Structure

This paper contains 9 sections, 21 theorems, 39 equations, 1 figure, 1 table, 1 algorithm.

Key Result

Theorem 1.1

(Theorem 2.2, climent:1999) Let $A = U-V$ be a $K$-weak regular splitting of type I (or type II). Then, $A^{-1}$ exists and $A^{-1}\geq_K 0$ if and only if $\rho(U^{-1}V)<1$.

Figures (1)

  • Figure 1: Monotonically increasing and decreasing convergence pattern of the iterative solution to the exact solution from two different initial approximated vector, which are nonnegative.

Theorems & Definitions (35)

  • Theorem 1.1
  • Theorem 2.1
  • Theorem 2.2
  • Lemma 2.3
  • Theorem 2.4
  • Theorem 2.5
  • Theorem 2.6
  • Theorem 3.1
  • Lemma 3.2
  • proof
  • ...and 25 more