Skip to main content
QUICK REVIEW

[Paper Review] Comparison results for semilinear elliptic equations using a new symmetrization method

François Hamel, Emmanuel Russ|arXiv (Cornell University)|Jan 8, 2014
Nonlinear Partial Differential Equations29 references3 citations
TL;DR

This paper introduces a novel symmetrization method for second-order semilinear elliptic equations, proving pointwise comparison results between solutions in a general domain Ω and radially symmetric solutions in the equimeasurable ball Ω*. The method symmetrizes the second-order terms and establishes that the solution in Ω is pointwise dominated by the solution in Ω* under natural constraints on coefficients and gradient growth (linear or quadratic).

ABSTRACT

In this paper, we prove some pointwise comparison results between the solutions of some second-order semilinear elliptic equations in a domain $Ω$ of $\R^n$ and the solutions of some radially symmetric equations in the equimeasurable ball $Ω^*$. The coefficients of the symmetrized equations in~$Ω^*$ satisfy similar constraints as the original ones in~$Ω$. We consider both the case of equations with linear growth in the gradient and the case of equations with at most quadratic growth in the gradient. Lastly, we show some improved quantified comparisons when the original domain is not a ball. The method is based on a symmetrization of the second-order terms.

Motivation & Objective

  • To establish pointwise comparison results between solutions of semilinear elliptic equations in a bounded C² domain Ω and corresponding radially symmetric solutions in the equimeasurable ball Ω*.
  • To develop a new symmetrization technique that preserves structural constraints of the original equation’s coefficients and gradient dependence.
  • To extend classical comparison results to equations with linear or at most quadratic growth in the gradient, including cases where the domain is not a ball.
  • To quantify improvements in comparison estimates when the original domain Ω is not radially symmetric, leveraging symmetrized second-order terms.
  • To recover and generalize Talenti’s seminal results on symmetrization and comparison via a new analytical framework.

Proposed method

  • The method introduces a symmetrization of the second-order elliptic operator by transforming the matrix field A(x) into a radially symmetric structure in Ω*, preserving the ellipticity and lower bound constraints.
  • It applies rearrangement techniques to the data: the solution u in Ω is rearranged to u* in Ω*, and the source term f is symmetrized to f*.
  • The symmetrized equation in Ω* takes the form −Δv + ṽα·∇v = f* in Ω*, with v = 0 on ∂Ω*, where ṽα is derived from the original gradient coefficient.
  • The proof relies on weak solutions in H¹₀(Ω) and H¹₀(Ω*), using compactness and convergence arguments (e.g., weak and a.e. convergence of symmetrized sequences).
  • It employs the maximum principle and comparison principles to show that the symmetrized solution v dominates the rearranged solution u* a.e. in Ω*.
  • The framework handles both linear and quadratic gradient growth by constructing auxiliary supersolutions and using boundedness and regularity of coefficients.

Experimental results

Research questions

  • RQ1Can a new symmetrization method be developed to compare solutions of semilinear elliptic equations in non-symmetric domains to radially symmetric solutions in equimeasurable balls?
  • RQ2How can the symmetrization process preserve the structural constraints of the original equation, particularly for coefficients with linear or quadratic gradient dependence?
  • RQ3What improvements in comparison estimates can be achieved when the original domain Ω is not a ball, and how can these be quantified?
  • RQ4To what extent can the proposed method recover or generalize Talenti’s classical comparison results for symmetric problems?
  • RQ5Under what conditions does the symmetrized solution in Ω* dominate the rearranged solution from Ω in the pointwise a.e. sense?

Key findings

  • The solution u in the domain Ω satisfies |u|* ≤ v a.e. in Ω*, where v is the unique H¹₀(Ω*) solution of the symmetrized equation −Δv + ṽα·∇v = |f|* in Ω* with v = 0 on ∂Ω*.
  • The symmetrized equation in Ω* preserves the radial symmetry of the coefficients and the structural constraints of the original problem, including the uniform ellipticity and lower bounds on the matrix field.
  • For equations with linear gradient growth, the method yields a comparison result via the symmetrized equation −Δv + ṽα·∇v = f* in Ω*, where ṽα is derived from the original coefficient field.
  • When the original domain Ω is not a ball, the method provides improved quantitative comparison estimates, particularly when min_Ω̅ b_Ω > 0, leading to stronger domination of u* by v.
  • The convergence of symmetrized sequences (u_k*, v_k*) is established in L¹(Ω*) and a.e. in Ω*, ensuring stability and consistency of the comparison under approximation.
  • The method recovers Talenti’s results as a special case when the original equation is linear and the domain is radially symmetric, confirming consistency with classical theory.

Better researchstarts right now

From reading papers to final review, dramatically reduce your research time.

No credit card · Free plan available

This review was created by AI and reviewed by human editors.