Quantitative homogenization of elliptic equations with infinitely many scales

Zhongwei Shen, Yao Xu, Jinping Zhuge · 2026

Read Quantitative homogenization of elliptic equations with infinitely many scales by Zhongwei Shen, Yao Xu, Jinping Zhuge (2026) 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-2605-02561.txt.

Introduction

Zhongwei Shen, Yao Xu, and Jinping Zhuge develop homogenization theory for elliptic equations whose coefficients oscillate at infinitely many scales at once. Homogenization replaces rapid wiggles with an effective smooth equation. Problems with endless scales arise in fractal materials and in diffusion in fluids, where structure repeats at finer and finer levels without end.

Under assumptions that the scales stay separated, the authors prove a qualitative homogenization theorem with optimal convergence rates in the mean-square sense. They also establish interior and boundary Lipschitz estimates that hold uniformly across all scale choices. Those uniform bounds are the technical core: they keep control of solutions no matter how many scales are active. The work gives a general frame for multi-scale media with no smallest scale.

Catalog record

CREATORZhongwei Shen, Yao Xu, Jinping Zhuge
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE148 KB
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2605.02561

Browse the full library · About the Tower of Babel

Full text

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

Quantitative homogenization of elliptic equations with infinitely many scales Zhongwei Shen, Yao Xu, Jinping Zhuge Abstract . In this paper, we develop a general homogenization theory for elliptic equations with coefficients that oscillate periodically at infinitely many scales ε = ( ε 1 , ε 2 , ⋯ ) ∈ ( 0 , 1 ) ∞ \varepsilon=(\varepsilon_{1},\varepsilon_{2},\cdots)\in(0,1)^{\infty} , with ε 1 > ε 2 > ⋯ \varepsilon_{1}>\varepsilon_{2}>\cdots and ε n → 0 \varepsilon_{n}\to 0 as n → ∞ n\to\infty . Such problems arise naturally in the study of fractal materials and diffusion in fluids. Under suitable scale-separation assumptions, we prove a qualitative homogenization theorem and obtain optimal L 2 L^{2} convergence rates. We also establish interior and boundary Lipschitz estimates that are uniform in ε \varepsilon . Keywords: Multiscale homogenization; convergence rates; uniform Lipschitz estimates. 1. Introduction 1.1. Motivations In this paper, we study the multiscale homogenization of elliptic equations. Precisely, we consider the boundary value problem, (1.1) \begin{cases}-\mathrm{div}(A^{\varepsilon}(x)\nabla u_{\varepsilon})=f&\text{in }\Omega,\\ u_{\varepsilon}=g&\text{on }\partial\Omega,\end{cases} where \Omega is a bounded Lipschitz domain in \mathbb{R}^{d} . In the classical homogenization theory of finitely many oscillating scales, one takes the coefficient matrix in the form of A^{\varepsilon}(x)=A(x/\varepsilon_{1},x/\varepsilon_{2},\cdots,x/\varepsilon_{n}) with \varepsilon=(\varepsilon_{1},\varepsilon_{2},\cdots,\varepsilon_{n})\in(0,1)^{n} and A=A(y_{1},y_{2},\cdots,y_{n}) satisfying certain self-similar structure conditions in each y_{i}\in\mathbb{R}^{d} . Here n is a fixed integer that represents the number of oscillating scales. The standard method for the multiscale homogenization problem is the so-called reiterated homogenization [ BLP78 , AB96 , LLPW01 , NSX20 ] . More recently, new approaches have been developed to study uniform regularity [ NZ23 , NZar ] and optimal convergence rates [ NXZar ] . Nevertheless, all of these results are proved using some inductive arguments on the number of scales n , and the constants in the estimates generally blow up as n\to\infty . However, in many theoretical treatments of physical problems, it is natural and crucial to consider fractal-like materials with a large number of or even infinitely many oscillating scales. Particularly in fluid dynamics it is believed that a large collection of vortices across a wide range of scales may be an important factor driving the turbulence evolution, as Lewis F. Richardson summarized in rhyming verse [ Ric07 ] back in 1922: “ Big whirls have little whirls that feed on their velocity, and little whirls have lesser whirls and so on to viscosity. ’’ While the real turbulence dynamics is much more complicated, the multiscale (or perpetual) homogenization theory has been applied in some simplified physical models, in the spirit of renormalization group method 1 1 1 Roughly speaking, renormalization group is a mathematical procedure to average out a material gradually from small scales to large scales, throw out the unnecessary small details and keep the most important information to get the large-scale properties of the material. , to understand phase transitions, turbulent mixing, anomalous diffusion, etc; see [ Ave96 ] for a nice introduction in this direction. We refer to [ AM91 , FP94 , MK99 , BAO02 , BAO03 ] and references therein for some early developments on this topic. Recently, reiterated homogenization with infinitely many scales has been applied to prove the anomalous dissipation for advection-diffusion equations [ AV25 , BSJW23 , BSJW26 ] ; and the coarse-graining theory 2 2 2 The coarse-graining theory in stochastic homogenization [ AS16 , AKM19 ] , developed by S. Armstrong, T. Kuusi, et al., can be thought of as a rigorous renormalization group argument. developed in stochastic homogenization has been used to obtain the superdiffusion for the same type of equations [ ABRK24 , ABRK26 ] . Due to the important applications described above, there is a practical need to develop a general homogenization theory for elliptic equations and systems 3 3 3 The methods in this paper do not distinguish between elliptic equations and systems. with infinitely many scales — this is the main objective of the present paper. At the same time, regarding multiscale homogenization itself, we also aim to understand the scope and potential limitations of homogenization problems involving infinitely many scales. 1.2. Set-ups and main results Returning to equation ( 1.1 ), the first question concerns the well-definedness of coefficients involving infinitely many scales. Let \varepsilon=(\varepsilon_{1},\varepsilon_{2},\cdots)\in(0,1)^{\infty} be the oscillating scales and 1=\varepsilon_{0}>\varepsilon_{1}>\varepsilon_{2}>\cdots with \lim_{n\to\infty}\varepsilon_{n}=0 . We define the coefficient matrix A^{\varepsilon}(x) with infinitely many scales as the L^{\infty} -limit of coefficient matrices with finitely many scales; i.e., A^{\varepsilon}(x)=\lim_{n\to\infty}A_{n}^{\varepsilon}(x) , where A_{n}^{\varepsilon}(x)=A_{n}(x,x/\varepsilon_{1},\cdots,x/\varepsilon_{n}) is a coefficient matrix with n oscillating scales, and A_{n}=A_{n}(y_{0},y_{1},\cdots,y_{n})\in C(\Omega\times\mathbb{T}^{d\times n}) is periodic in each y_{i}\in\mathbb{T}^{d} for 1\leq i\leq n . If A_{n}^{\varepsilon}(x) is continuous for each n , then the L^{\infty} -limit A^{\varepsilon}(x) is also continuous and well-defined. Moreover, if we set \displaystyle B_{n}(y_{0},y_{1},\cdots,y_{n})=A_{n}(y_{0},y_{1},\cdots,y_{n})-A_{n-1}(y_{0},y_{1},\cdots,y_{n-1}), then (1.2) \displaystyle A_{n}(y_{0},y_{1},\cdots,y_{n})=\sum_{\ell=0}^{n}B_{\ell}(y_{0},y_{1},\cdots,y_{\ell}). Thus, in this paper, we shall consider A_{n} in a form of ( 1.2 ) and (1.3) A^{\varepsilon}(x)=\lim_{n\to\infty}A_{n}^{\varepsilon}(x)\quad\text{uniformly in }x\in\Omega. Moreover, we assume that each d\times d matrix B_{\ell}(y_{0},y_{1},\cdots,y_{\ell}) , defined on \Omega\times\mathbb{R}^{d\times\ell} , is continuous and periodic in y_{1},\cdots,y_{\ell} . Before stating our main assumptions and results, let us first mention that the coefficient matrix A^{\varepsilon}(x) defined above has a fractal structure with infinitely many scales. Indeed, we can formally write A_{\infty}(y_{0},y_{1},\cdots)=\sum_{\ell=0}^{\infty}B_{\ell}(y_{0},y_{1},\cdots,y_{\ell}), and A^{\varepsilon}(x)=A_{\infty}\Big(x,\frac{x}{\varepsilon_{1}},\frac{x}{\varepsilon_{2}},\cdots\Big). At scale \varepsilon_{0}=1 , we see infinitely many oscillating scales in a decreasing order \varepsilon_{1}>\varepsilon_{2}>\cdots . If we zoom in and look at the coefficients at the scale \varepsilon_{m} , the rescaled coefficient matrix takes a form of A^{\varepsilon}(\varepsilon_{m}x)=A_{\infty}\Big(\varepsilon_{m}x,\frac{x}{\varepsilon_{1}/\varepsilon_{m}},\frac{x}{\varepsilon_{2}/\varepsilon_{m}},\cdots,\frac{x}{\varepsilon_{m-1}/\varepsilon_{m}},x,\frac{x}{\varepsilon_{m+1}/\varepsilon_{m}},\cdots\Big). Observe that although the first m oscillating scales are flattened, we are still left with infinitely many oscillating scales: \frac{\varepsilon_{m+1}}{\varepsilon_{m}}>\frac{\varepsilon_{m+2}}{\varepsilon_{m}}>\cdots . This means that no matter how small a scale we zoom in to, we always see the oscillations of coefficients at infinitely many scales — a phenomenon that demonstrates the characteristic of fractal structures. We now introduce the ellipticity, boundedness and periodicity assumptions for \{A_{n}\}_{n\in\mathbb{N}} given in ( 1.2 ). • Ellipticity: There exists a fixed \mu>0 such that every n\in\mathbb{N} , (1.4) \xi\cdot A_{n}\xi\geq\mu|\xi|^{2}\quad\text{for any }\xi\in\mathbb{R}^{d}; • Boundedness: (1.5) \sum_{\ell=0}^{\infty}\|B_{\ell}\|_{L^{\infty}}\leq\mu^{-1}. • 1-Periodicity: For arbitrarily n\in\mathbb{N}_{+}:=\mathbb{N}\setminus\{0\} , (1.6) A_{n}(y_{0},y_{1}+z_{1},\cdots,y_{n}+z_{n})=A_{n}(y_{0},y_{1},\cdots,y_{n}), for any z_{j}\in\mathbb{Z}^{d},j=1,2,\cdots,n . Clearly, ( 1.5 ) implies that each A_{n} satisfies the boundedness condition uniformly in n , i.e., (1.7) |A_{n}\xi|\leq\mu^{-1}|\xi|\quad\text{for any }\xi\in\mathbb{R}^{d}. As a result, the limiting matrix A^{\varepsilon}(x) also satisfies the same ellipticity and boundedness conditions and the Dirichlet problem ( 1.1 ) is well-posed for f\in H^{-1}(\Omega) and g\in H^{1/2}(\partial\Omega) . Our first result is a qualitative homogenization theorem. The same as in homogenization with finitely many scales [ AB96 ] , the qualitative homogenization holds under the scale-separation condition: (1.8) \varepsilon_{1}\to 0\text{ and }\frac{\varepsilon_{k+1}}{\varepsilon_{k}}\to 0\quad\text{for any }k\geq 1, where \varepsilon=(\varepsilon_{1},\varepsilon_{2},\cdots) is understood as a sequence of elements in (0,1)^{\infty} without explicit indices. Throughout the paper, for convenience, we will write the condition ( 1.8 ) simply as \varepsilon\mathrel{\mathchoice{\rightarrow\mspace{-10mu}\rightarrow}{\rightarrow\mspace{-10mu}\rightarrow}{\rightarrow\mspace{-10mu}\rightarrow}{\rightarrow\mspace{-10mu}\rightarrow}}0 . Theorem 1.1 . Let \Omega be a bounded Lipschitz domain in \mathbb{R}^{d} . Assume ( 1.4 ), ( 1.5 ) and ( 1.6 ). Let f\in H^{-1}(\Omega) and g\in H^{1/2}(\partial\Omega) . Then the solution u_{\varepsilon} of ( 1.1 ) converges to u_{0} weakly in H^{1}(\Omega) as \varepsilon\mathrel{\mathchoice{\rightarrow\mspace{-10mu}\rightarrow}{\rightarrow\mspace{-10mu}\rightarrow}{\rightarrow\mspace{-10mu}\rightarrow}{\rightarrow\mspace{-10mu}\rightarrow}}0 , where u_{0} is the solution to (1.9) \begin{cases}-\mathrm{div}(\widehat{A}\nabla u_{0})=f&\text{in }\Omega,\\ u_{0}=g&\text{on }\partial\Omega,\end{cases} and \widehat{A}=\widehat{A}(x) is the homogenized coefficient matrix given in Theorem 2.5 . Moreover, as \varepsilon\mathrel{\mathchoice{\rightarrow\mspace{-10mu}\rightarrow}{\rightarrow\mspace{-10mu}\rightarrow}{\rightarrow\mspace{-10mu}\rightarrow}{\rightarrow\mspace{-10mu}\rightarrow}}0 , A^{\varepsilon}(x) H -converges to

Read the complete text