Francisco J. Fernández,F. Adrián F. Tojo
Abstract:This work is devoted to the obtaining of a new numerical scheme based in quadrature formulas for the Lebesgue-Stieltjes integral for the approximation of Stieltjes ordinary differential equations. This novel method allows us to numerically approximate models based in Stieltjes ordinary differential equations for which no explicit solution is known. We prove several theoretical results related to the consistency, convergence and stability of the numerical method. We also obtain the explicit solution of the Stieltjes linear ordinary differential equation, and we use it to validate the numerical method. Finally, we present some numerical results that we have obtained for a realistic population model based in a Stieltjes differential equation.
What problem does this paper attempt to address?
The problem that this paper attempts to solve is to develop a new numerical method for approximating the solutions of Stieltjes ordinary differential equations (SODEs). Specifically, the author proposes a new numerical scheme based on the quadrature formula of the Lebesgue - Stieltjes integral to numerically approximate Stieltjes ordinary differential equation models for which no known explicit solutions exist.
### Problem Background
Stieltjes ordinary differential equations are a class of generalized differential equations, and the definition of their derivatives involves a left - continuous, non - decreasing function \( g \). Such equations are very important in many practical applications, such as in the fields of population dynamics and biomathematics. However, for most practical problems, it is difficult to obtain the analytical solutions of these equations, so effective numerical methods are required for approximate solutions.
### Main Contributions of the Paper
1. **New Numerical Scheme**: The author proposes a new numerical method based on the quadrature formula of the Lebesgue - Stieltjes integral, which can effectively approximate the solutions of Stieltjes ordinary differential equations.
2. **Theoretical Analysis**: The paper proves the consistency, convergence, and stability of this numerical method, ensuring the reliability of the method.
3. **Verification and Application**: The author verifies the effectiveness of the numerical method by solving the explicit solutions of linear Stieltjes ordinary differential equations and applies it to a real silkworm population model based on Stieltjes differential equations, demonstrating the practical application value of this method.
### Mathematical Description
Consider the Stieltjes ordinary differential equation in the following form:
\[
\begin{cases}
x'_g(t)=f(t, x(t)), & \text{for almost all } t\in[0, T), \\
x(0) = x_0,
\end{cases}
\]
where \( x'_g(t) \) is the Stieltjes derivative with respect to \( g \), defined as:
\[
x'_g(t)=
\begin{cases}
\lim_{s\rightarrow t}\frac{x(s)-x(t)}{g(s)-g(t)}, & \text{if } t\notin D_g, \\
\frac{x(t^+)-x(t)}{g(t^+)-g(t)}, & \text{if } t\in D_g,
\end{cases}
\]
where \( D_g \) represents the set of discontinuity points of \( g \).
### Core Idea of the Numerical Method
To approximate the solutions of the above equations, the author uses the quadrature formula based on the Lebesgue - Stieltjes integral. Specifically, they decompose the integral term into a continuous part and a jump part and approximate them separately. In this way, a predictor - corrector method can be obtained, which can provide more accurate solutions within each time step.
### Conclusion
This paper proposes a novel and effective numerical method for approximating the solutions of Stieltjes ordinary differential equations. Through strict theoretical analysis and practical application verification, the effectiveness and reliability of this method are proved. This provides a powerful tool for solving complex problems involving Stieltjes derivatives.