Harnack-type inequalities for nonlinear evolution equations
Jessica Slegers · 2026
Read Harnack-type inequalities for nonlinear evolution equations by Jessica Slegers (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-2601-05642.txt.
Introduction
Jessica Slegers surveys, in a solo 2026 monograph, Harnack-type inequalities for nonlinear evolution equations. Harnack inequalities bound how much an answer of a partial differential equation can vary, giving qualitative control over its behavior. They began with harmonic functions and spread to wider elliptic and parabolic problems.
Slegers follows methods from Li and Yau for the linear heat equation, later extended by Auchmuty and Bao to nonlinear cases. A minimization problem that appears naturally in the proofs is studied first. Three general results are then applied to the porous medium equation and to a p-Laplace diffusion problem. A closing application proves local Holder continuity of answers using Moser's standard arguments.
Catalog record
| CREATOR | Jessica Slegers |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 124 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2601-05642
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Harnack-type inequalities for nonlinear evolution equations Jessica Slegers Abstract A thesis submitted in partial fulfilment of the requirements for the degree of B.Sc. (Honours) School of Mathematics and Statistics University of Sydney Australia November 2021 Preface Harnack inequalities are useful qualitative tools for understanding the properties of partial differential equations. Originally discovered as a property of harmonic functions, Harnack inequalities have since been studied for solutions of wider classes of elliptic and parabolic problems. In this monograph, we take particular interest in deriving Harnack inequalities for solutions of nonlinear evolution equations. We focus on exploring the methods introduced by Li and Yau in the case of the linear heat equation and later extended to nonlinear problems by Auchmuty and Bao. Prior to presenting these results, we study a minimisation problem, which appears naturally in the proofs. After establishing a family of three general Harnack inequality results by Auchmuty and Bao, we investigate applications to deriving Harnack inequalities satisfied by solutions of the porous medium equation and weak solutions of the parabolic problem associated with the p -Laplace operator, which we refer to here as the p -diffusion equation. Finally, we demonstrate a common application of Harnack inequalities by proving the local space-time Hölder continuity of solutions to a class of linear evolution problems. The proof is based on methods introduced by Moser during his seminal work on Harnack inequalities during the 1960s. We conclude by suggesting potential opportunities for future work following on from the topics discussed here. Chapter 1 Introduction When studying partial differential equations, it is important to develop qualitative and quantitative tools, such as inequalities, which can be used to discover features of their solutions, for example, their Hölder continuity or maximum principles. In 1887, Harnack [ 21 ] proved the following inequality for positive harmonic functions defined on the open ball B_{R}(x_{0})\subseteq\mathbb{R}^{2} : \frac{R-r}{R+r}u(x_{0})\leq u(x)\leq\frac{R+r}{R-r}u(x_{0}). This inequality holds for all x\in B_{r}(x_{0}) and r<R . We note that throughout this monograph, a function u defined on \Omega will be called positive if u(x)>0 for a.e. x\in\Omega and nonnegative if u(x)\geq 0 for a.e. x\in\Omega . An analogous formulation of this result may also be given when u is a positive harmonic function on a domain \Omega\subseteq\mathbb{R}^{d} for d\geq 1 . Often, Harnack’s inequality is stated in the following form \gamma^{-1}\sup_{B_{r}(x_{0})}u\leq u(x_{0})\leq\gamma\inf_{B_{r}(x_{0})}u, (1.1) which is due to Kellogg [ 23 ] . Here, we assume that \bar{B}_{r}(x_{0})\subseteq\Omega and importantly, the constant \gamma>1 is independent of the function u . Alternatively, the ball B_{r}(x_{0}) may be replaced with any subdomain \Omega^{\prime} compactly contained in \Omega , meaning that \overline{\Omega^{\prime}}\subseteq\Omega . Harnack’s inequality has several important consequences in the theory of harmonic functions. (i) (Liouville’s Theorem, [ 17 ] ) Every nonnegative harmonic function on \mathbb{R}^{d} is constant. (ii) (Removable Singularity Theorem, [ 6 ] ) Let d\geq 3 . If a function u:B_{r}(0)\setminus\{0\}\rightarrow\mathbb{R} is harmonic and u(x)=o(|x|^{2-d}) for x\rightarrow 0 , then u(0) can be defined so that u:B_{r}(0)\rightarrow\mathbb{R} is a harmonic function. (iii) (Harnack’s First Convergence Theorem, [ 19 ] ) If a sequence of harmonic functions on a bounded domain \Omega\subset\mathbb{R}^{d} , which are continuous on \bar{\Omega} , converges uniformly on the boundary \partial\Omega , then the sequence converges uniformly in \Omega to a harmonic function. (iv) (Harnack’s Second Convergence Theorem, [ 19 ] ) If a sequence (u_{n}) of harmonic functions on a connected open set \Omega\subset\mathbb{R}^{d} is monotone increasing and there exists x_{0}\in\Omega such that (u_{n}(x_{0})) is convergent, then (u_{n}) converges uniformly on every compact subset of \Omega to a harmonic function u . Following Harnack’s discovery, attention has been given to proving similar inequalities for parabolic equations. In the parabolic case, a time dependence can be seen in the resulting inequalities. In 1954, Hadamard [ 20 ] and Pini [ 31 ] independently derived Harnack-type inequalities of the form \gamma^{-1}\sup_{x\in B_{r}(x_{0})}u(x,t_{0}-r^{2})\leq u(x_{0},t_{0})\leq\gamma\inf_{x\in B_{r}(x_{0})}u(x,t_{0}+r^{2}) (1.2) for positive solutions to the heat equation, u_{t}=\Delta u\qquad\text{in }\Omega\times(0,\infty). (1.3) In inequality ( 1.2 ), we see a waiting time from t_{0}-r^{2} to t_{0} and from t_{0} to t_{0}+r^{2} . Using the physical interpretation of ( 1.3 ) as describing the distribution of heat in a medium, this means that if the temperature u is known at a point (x_{0},t_{0}) , then after waiting some time, the temperature everywhere in a neighbourhood of x_{0} will be at least \frac{u(x_{0},t_{0})}{\gamma} . This waiting time is in fact necessary for the result to hold, as has been demonstrated using counterexamples, such as the following one contributed by Moser [ 29 ] . Suppose there exists \gamma>1 independent of u such that \sup_{x\in K}u(x,t_{0})\leq\gamma\inf_{x\in K}u(x,t_{0}) (1.4) for all solutions u of u_{t}=u_{xx}\qquad\text{in }\mathbb{R}\times(0,\infty), (1.5) where K\subset\subset\mathbb{R} and t_{0}>0 is fixed. The function u_{\xi}(x,t):=t^{-1/2}e^{-(x+\xi)^{2}/4t} is known to be a solution of ( 1.5 ) for all \xi\in\mathbb{R} . Observe that for any x_{0}>0 fixed, one has that \lim_{\xi\rightarrow-\infty}\frac{u_{\xi}(0,1)}{u_{\xi}(x_{0},1)}=\lim_{\xi\rightarrow-\infty}e^{x_{0}^{2}/4}e^{x_{0}\xi/4}=0. Applying ( 1.4 ) for some K\subset\subset\mathbb{R} containing x_{0} and 0 yields \left(\frac{u_{\xi}(0,1)}{u_{\xi}(x_{0},1)}\right)^{-1}\leq\gamma. Taking \xi\rightarrow-\infty , we reach a contradiction, since the left-hand side becomes unbounded in the limit. In 1964, Moser [ 29 ] also expanded on this work by obtaining a similar result for weak solutions of more general parabolic differential equations with variable coefficients, that is, equations in the following divergence form \frac{\partial u}{\partial t}=\sum_{k,l=1}^{n}\frac{\partial}{\partial x_{k}}\left(a_{kl}(x,t)\frac{\partial u}{\partial x_{l}}\right)\qquad\text{in }\Omega\times(0,T). (1.6) Here, the coefficients a_{kl}(x,t) are measurable functions with a_{kl}=a_{lk} that satisfy the uniform ellipticity condition \lambda^{-1}\leq\sum_{k,l=1}^{n}a_{kl}(x,t)\xi_{k}\xi_{l}\leq\lambda (1.7) for all \xi\in\mathbb{R}^{d} with \sum_{k}\xi^{2}_{k}=1 . In particular, Moser proved that there exists a positive constant C such that \sup_{B_{r}(x_{0})\times(t_{1}^{-},t_{2}^{-})}u\leq C\inf_{B_{r}(x_{0})\times(t_{1}^{+},t_{2}^{+})}u for all nonnegative weak solutions u of ( 1.6 ) where r>0 is such that \bar{B}_{r}(x_{0})\subseteq\Omega and 0<t_{1}^{-}<t_{2}^{-}<t_{1}^{+}<t_{2}^{+}<T . Here, one must understand the supremum and infimum in the sense of the essential supremum and essential infimum respectively. Moser then used the oscillation of weak solutions of ( 1.6 ) to prove that these are space-time Hölder continuous on the interior of a rectangular domain. Although Hölder continuity had already been established by Nash [ 30 ] in 1958, Moser presented a new method for proving this result, which we will explore in Chapter 5 . Research in pursuit of Harnack-type inequalities has been heavily influenced by the work of Li and Yau [ 27 ] , who developed new methods for deriving such results. In their analysis of positive solutions of the heat equation ( 1.3 ) on \mathbb{R}^{d}\times(0,T) , the sharp gradient estimate \frac{|\nabla u|^{2}}{u^{2}}-\frac{u_{t}}{u}\leq\frac{d}{2t} (1.8) was found. By integrating this inequality over a path connecting (x_{1},t_{1}) and (x_{2},t_{2}) in \mathbb{R}^{d}\times(0,T) with 0<t_{1}<t_{2}<T , it follows that for all such points u(x_{2},t_{2})\geq u(x_{1},t_{1})\left(\frac{t_{1}}{t_{2}}\right)^{d/2}e^{-\tfrac{|x_{2}-x_{1}|^{2}}{4(t_{2}-t_{1})}}. (1.9) We remark that Li and Yau’s results have particular significance, since they hold more generally on complete Riemannian manifolds with nonnegative Ricci curvature. However, the exploration of Harnack inequalities on manifolds is beyond the scope of this monograph and we instead restrict our attention to the Euclidean space \mathbb{R}^{d} . Harnack-type estimates analogous to ( 1.2 ) have also been formulated and proven for parabolic and elliptic equations in non-divergence form, for example, by Krylov and Safonov [ 25 , 32 ] . Similar to the case of operators in divergence form, this result can be used to prove the Hölder continuity of the solutions. Another topic