Grey system model on time scales

Wanli Xie · 2025

Read Grey system model on time scales by Wanli Xie (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2501-09358.txt.

Introduction

Wanli Xie merges grey system theory with the calculus of time scales, in a 2025 paper. Grey system theory models systems where information is uncertain or incomplete, a common state in real data. Time scales offer one framework that covers both continuous time and discrete time, rather than treating them as separate worlds.

The merged model, called the Grey System Model on Time Scales, is offered as a robust tool for hybrid systems. Such systems mix events that unfold smoothly with events that jump at set points, so their time domains vary. Xie presents the construction as a way to analyze those mixed cases inside one model instead of two.

Catalog record

CREATORWanli Xie
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE11 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2501-09358

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Grey system model on time scales Wanli Xie Abstract The Grey System Theory (GST) is a powerful mathematical framework employed for modeling systems with uncertain or incomplete information. This paper proposes an integration of the GST with time scales, a generalized approach that encompasses both discrete and continuous time models. The proposed model, called the Grey System Model on Time Scales (GST-T), offers a robust solution for analyzing hybrid systems where events occur on varying time domains. 1 Introduction In many real-world applications, systems exhibit uncertain, incomplete, or imprecise data, making it difficult to develop accurate models using traditional methods. In such cases, the Grey System Theory, proposed by Deng [ 1 ] in 1982, has proven to be a powerful tool for modeling, analyzing, and forecasting systems with sparse or uncertain data. GST focuses on the use of limited data to predict future trends, especially when full information about a system is not available. A core component of GST is the Grey Model (GM), particularly the GM(1,1) model, which has become a widely used method for time series forecasting [ 2 ] . The GM(1,1) model is a first-order differential equation that describes the evolution of a single variable within a system. It is particularly effective in addressing systems with incomplete historical data and has been successfully applied across various fields, such as economics, engineering, environmental sciences, and social sciences. However, in the equation \frac{{d{x^{(1)}}(t)}}{{dt}}+a{x^{(1)}}(t)=b, the variable x^{(1)}(t) is not rigorously defined but is instead characterized within the framework of difference equations. This transition between difference and differential equations introduces additional challenges, further limiting the generalizability of the model. To address this limitation, a more generalized approach is required. This is where time scales theory comes into play. Time scales theory provides a unifying framework for studying dynamic systems that evolve over both continuous and discrete time domains [ 3 ] . A time scale is defined as a closed subset of real numbers, which may consist of discrete, continuous, or a combination of both types of points. This generalization allows for the modeling of hybrid systems that exhibit both discrete events and continuous processes in a seamless manner. 2 Preliminaries and problem formulation 2.1 The time scales calculus A time scale \mathbb{T} is any nonempty closed subset of the real numbers \mathbb{R} . Examples include \mathbb{R} (real numbers), \mathbb{Z} (integers), \mathbb{N} (natural numbers), finite unions of closed intervals (e.g., [0,1]\cup[2,3] ), and the Cantor set. Sets that are not closed, such as \mathbb{Q} (rationals) or open intervals like (0,1) , are not time scales. A time scale \mathbb{T} inherits the standard topology from \mathbb{R} . Definition 1 ( [ 3 ] ) . Let \mathbb{T} be a time scale. The forward jump operator \sigma:\mathbb{T}\to\mathbb{T} is defined as \sigma(t):=\inf\{s\in\mathbb{T}\mid s>t\},\quad\forall t\in\mathbb{T}. The backward jump operator \rho:\mathbb{T}\to\mathbb{T} is given by \rho(t):=\sup\{s\in\mathbb{T}\mid s<t\},\quad\forall t\in\mathbb{T}. Definition 2 ( [ 3 ] ) . The graininess function \mu:\mathbb{T}\to[0,\infty) associated with a time scale \mathbb{T} is defined as \mu(t):=\sigma(t)-t,\quad\forall t\in\mathbb{T}. Definition 3 ( [ 3 ] ) . Let \mathbb{T} be a time scale. The set \mathbb{T}^{\kappa} is defined as \mathbb{T}^{\kappa}:=\begin{cases}\mathbb{T}\setminus\{M\},&\text{if }M=\sup\mathbb{T}\text{ and }M\text{ is left-scattered},\\ \mathbb{T},&\text{otherwise},\end{cases} where M=\sup\mathbb{T} represents the supremum of the time scale \mathbb{T} . Definition 4 ( [ 3 ] ) . Let f:\mathbb{T}\to\mathbb{R} be a function on a time scale \mathbb{T} , and let t\in\mathbb{T}^{\kappa} , where \mathbb{T}^{\kappa} is the set of points in \mathbb{T} excluding right-scattered maximums. The delta derivative of f at t , denoted f^{\Delta}(t) , is the unique value (if it exists) such that, for any \varepsilon>0 , there exists a neighborhood U of t satisfying \left|f(\sigma(t))-f(s)-f^{\Delta}(t)(\sigma(t)-s)\right|\leq\varepsilon|\sigma(t)-s|,\quad\forall s\in U. Definition 5 ( [ 3 ] ) . Let f:\mathbb{T}\to\mathbb{R} be a regulated function. A function F:\mathbb{T}\to\mathbb{R} is called a pre-antiderivative of f if it satisfies F^{\Delta}(t)=f(t),\quad\forall t\in\mathbb{T}^{\kappa}, where F^{\Delta}(t) denotes the delta derivative of F . Definition 6 ( [ 3 ] ) . Let f:\mathbb{T}\to\mathbb{R} be rd-continuous. Then there exists an antiderivative F:\mathbb{T}\to\mathbb{R} such that F(t)=\int_{t_{0}}^{t}f(\tau)\Delta\tau,\quad\forall t\in\mathbb{T} for any fixed t_{0}\in\mathbb{T} . Definition 7 ( [ 3 ] ) . A function p:\mathbb{T}\to\mathbb{R} is called \nu - regressive if 1-\nu(t)p(t)\neq 0,\quad\forall t\in\mathbb{T}^{\kappa}. The set of all \nu -regressive is denoted by \mathcal{R}_{\nu}=\{p:\mathbb{T}\to\mathbb{R}:p\text{ is ld-continuous and $\nu$-regressive}\}. Definition 8 ( [ 3 ] ) . Let h>0 , set Z_{h}:=\left\{z\in\mathbb{C}:-\frac{\pi}{h}<\operatorname{Im}(z)<\frac{\pi}{h}\right\},\quad C_{h}:=\left\{z\in\mathbb{C}:z\neq\frac{1}{h}\right\}. (1) Then the \nu -cylinder transformation \hat{\xi}_{h}:C_{h}\to Z_{h} is given by \hat{\xi}_{h}(z):=-\frac{1}{h}\operatorname{Log}(1-zh), (2) where \operatorname{Log} is the principal logarithm function. For the degenerate case h=0 , the transformation is defined as \hat{\xi}_{0}(z):=z for all z\in\mathbb{C}_{0}:=\mathbb{C} . Definition 9 ( [ 3 ] ) . Consider p\in\mathcal{R}_{\nu} , where \mathcal{R}_{\nu} denotes the set of appropriately defined functions. The nabla exponential function is then formally defined as \hat{e}_{p}(t,s):=\exp\left(\int_{s}^{t}\hat{\xi}_{\nu(\tau)}(p(\tau))\nabla\tau\right) (3) for all s,t\in\mathbb{T} . 2.2 Grey system model on time scales Let x(t) be the original data function defined on \mathbb{T} . The accumulated generating operation (AGO) is applied to the data as follows: y(t):=y(t_{0})+\int_{t_{0}}^{t}x(\tau)\Delta\tau, For continuous time ( \mathbb{T}=\mathbb{R} ), the integral becomes: y(t)=y(t_{0})+\int_{t_{0}}^{t}x(\tau)d\tau. For discrete time ( \mathbb{T}=\mathbb{Z} ), the integral is replaced by a summation: y(t)=\sum_{\tau=t_{0}+1}^{t}x(\tau). The core equation for the GM(1,1) model on time scales is derived by replacing the regular derivative with the \Delta -derivative. The resulting equation can be defined as y^{\Delta}(t)+ay(t)=b. (4) The solution of equation ( 4 ) can be derived by y(t)=\hat{e}_{-a}(t,t_{0})\left[y(t_{0})-\frac{b}{a}\right]+\frac{b}{a}. For continuous time ( T=\mathbb{R} ), the solution of the model takes the form e_{-a}(t,t_{0})=e^{-a(t-t_{0})}\left[y(t_{0})-\frac{b}{a}\right]+\frac{b}{a} , which is consistent with the solution of the classical GM(1,1) model. For discrete time ( T=\mathbb{Z} ), the solution is expressed as e_{-a}(t,t_{0})=(1-a)^{t-t_{0}}\left[y(t_{0})-\frac{b}{a}\right]+\frac{b}{a} . 3 Applications and discussion The classical grey model can be constructed in both continuous and discrete time domains. It is possible to establish the model exclusively in the continuous domain or solely in the discrete domain. For instance, the Riemann-Liouville (RL) integral can be used to construct a continuous operator, and its solution can be represented in a numerical form [ 4 ] . The RL integral is defined as D_{0,t}^{-r}f(t)=\frac{1}{\Gamma(r)}\int_{0}^{t}(t-s)^{r-1}f(s)\,ds,\quad r>0, and its numerical form is given by \displaystyle D_{0,t}^{-r}f\left(t_{n}\right) \displaystyle=\frac{1}{\Gamma(r)}\sum_{k=0}^{n-1}\int_{t_{k}}^{t_{k+1}}\left(t_{n}-s\right)^{r-1}f\left(t_{k}\right)ds (5) \displaystyle=\frac{-1}{\Gamma(r)}\sum_{k=0}^{n-1}\int_{t_{k}}^{t_{k+1}}\left(t_{n}-s\right)^{r-1}d\left(t_{n}-s\right)f\left(t_{k}\right) \displaystyle=\frac{-1}{\Gamma(r)}\sum_{k=0}^{n-1}\frac{1}{r}\left[\left(t_{n}-t_{k+1}\right)^{r}-\left(t_{n}-t_{k}\right)^{r}\right]f\left(t_{k}\right) \displaystyle=\frac{-\Delta t^{r}}{\Gamma(r+1)}\sum_{k=0}^{n-1}-\left[(n-k)^{r}-(n-k-1)^{r}\right]f\left(t_{k}\right) \displaystyle=\frac{\Delta t^{r}}{\Gamma(r+1)}\sum_{k=0}^{n-1}\left[(n-k)^{r}-(n-k-1)^{r}\right]f\left(t_{k}\right) \displaystyle=\frac{\Delta t^{r}}{\Gamma(r+1)}\sum_{k=1}^{n}\left[(n+1-k)^{r}-(n+1-k-1)^{r}\right]f\left(t_{k}\right). This numerical form can also be expressed in matrix form as D=\left[{\begin{array}[]{*{20}{c}}{\Psi(1)}&{\Psi(1)}&{\Psi(1)}&\cdots&{\Psi(1)}\\ 0&{\Psi(2)}&{\Psi(2)}&\cdots&{\Psi(2)}\\ 0&0&{\Psi(3)}&\cdots&{\Psi(3)}\\ \vdots&\vdots&\vdots&\ddots&\vdots\\ 0&0&0&\cdots&{\Psi(n)}\end{array}}\right], (6) where \Psi(k)=\frac{1}{{\Gamma(r+1)}}\left[{{{(n+1-k)}^{r}}-{{(n+1-k-1)}^{r}}}\right]. (7) Figure 1: Fractional integral comparison for different orders. The inverse matrix of D can be used to compute predictions based on the grey model, providing an efficient approach for obtaining forecasted values. In this study, we examine the fractional integral of the power function y(t)=t^{2} using both analytical and numerical approaches to validate the efficacy of the proposed methods. The analytical solution is derived based on the RL integral, yielding I_{t}^{r}y(t)=\frac{\Gamma(3)}{\Gamma(3+r)}t^{2+r} , where \Gamma(\cdot) denotes the Gamma function and r represents the fractional order. To complement the analytical results, we employ two numerical techniques: the accumulative method [ 5 ] and the Riemann-Liouville (RL) numerical integration. The accumulative method computes the fractional cumulative sum by applying weighted coefficients derived from the Gamma function, while the RL numerical integration discretizes the integral using finite difference schemes to approximate the fractional integral. The experiment is conducted over the time interval t\in[1,15] with fractional orders r=0.4 and r=0.3 . Comparative analysis, as illustrated in Figure 1 , demonstrates

Read the complete text