We prove interpolating estimates providing a bound for the oscillation of a function in terms of two Lp norms of its gradient. They are based on a pointwise bound of a function on cones in terms of the Riesz potential of its gradient. The estimates hold for a general class of domains, including, e.g., Lipschitz domains. All the constants involved can be explicitly computed. As an application, we show how to use these estimates to obtain stability for Alexandrov's Soap Bubble Theorem and Serrin's overdetermined boundary value problem. The new approach results in several novelties and benefits for these problems.
Citation: Rolando Magnanini, Giorgio Poggesi. Interpolating estimates with applications to some quantitative symmetry results[J]. Mathematics in Engineering, 2023, 5(1): 1-21. doi: 10.3934/mine.2023002
[1] | YanYan Li . Symmetry of hypersurfaces and the Hopf Lemma. Mathematics in Engineering, 2023, 5(5): 1-9. doi: 10.3934/mine.2023084 |
[2] | Bruno Bianchini, Giulio Colombo, Marco Magliaro, Luciano Mari, Patrizia Pucci, Marco Rigoli . Recent rigidity results for graphs with prescribed mean curvature. Mathematics in Engineering, 2021, 3(5): 1-48. doi: 10.3934/mine.2021039 |
[3] | Filippo Gazzola, Gianmarco Sperone . Remarks on radial symmetry and monotonicity for solutions of semilinear higher order elliptic equations. Mathematics in Engineering, 2022, 4(5): 1-24. doi: 10.3934/mine.2022040 |
[4] | Michiaki Onodera . Linear stability analysis of overdetermined problems with non-constant data. Mathematics in Engineering, 2023, 5(3): 1-18. doi: 10.3934/mine.2023048 |
[5] | Qiang Guang, Qi-Rui Li, Xu-Jia Wang . Flow by Gauss curvature to the $ L_p $ dual Minkowski problem. Mathematics in Engineering, 2023, 5(3): 1-19. doi: 10.3934/mine.2023049 |
[6] | Hyeonbae Kang, Shigeru Sakaguchi . A symmetry theorem in two-phase heat conductors. Mathematics in Engineering, 2023, 5(3): 1-7. doi: 10.3934/mine.2023061 |
[7] | Nicolai Krylov . On parabolic Adams's, the Chiarenza-Frasca theorems, and some other results related to parabolic Morrey spaces. Mathematics in Engineering, 2023, 5(2): 1-20. doi: 10.3934/mine.2023038 |
[8] | Stefano Borghini . Symmetry results for Serrin-type problems in doubly connected domains. Mathematics in Engineering, 2023, 5(2): 1-16. doi: 10.3934/mine.2023027 |
[9] | Eleonora Cinti, Roberto Ognibene, Berardo Ruffini . A quantitative stability inequality for fractional capacities. Mathematics in Engineering, 2022, 4(5): 1-28. doi: 10.3934/mine.2022044 |
[10] | Italo Capuzzo Dolcetta . The weak maximum principle for degenerate elliptic equations: unbounded domains and systems. Mathematics in Engineering, 2020, 2(4): 772-786. doi: 10.3934/mine.2020036 |
We prove interpolating estimates providing a bound for the oscillation of a function in terms of two Lp norms of its gradient. They are based on a pointwise bound of a function on cones in terms of the Riesz potential of its gradient. The estimates hold for a general class of domains, including, e.g., Lipschitz domains. All the constants involved can be explicitly computed. As an application, we show how to use these estimates to obtain stability for Alexandrov's Soap Bubble Theorem and Serrin's overdetermined boundary value problem. The new approach results in several novelties and benefits for these problems.
Let Ω⊂RN be a bounded domain. For 1≤p≤∞ the number ‖f‖p,Ω, will denote the Lp-norm of a measurable function f:Ω→R with respect to the normalized Lebesgue measure dμx=dx/|Ω|.
In Theorems 2.4 and 2.7, for N<q≤∞, we prove the following interpolating inequalities, which hold true for any f∈W1,q(Ω):
max¯Ωf−min¯Ωf≤c{‖∇f‖p,Ωfor p>N‖∇f‖N,Ωlog(e‖∇f‖q,Ω/‖∇f‖N,Ω), for p=N,‖∇f‖αp,qp,Ω‖∇f‖1−αp,qq,Ω, for 1≤p<N. | (1.1) |
Here,
αp,q=p(q−N)N(q−p) for N<q<∞,αp,∞=pN. |
Notice that simply combining the Morrey-Sobolev embedding W1,r↪C0,1−N/r for r>N and the classical interpolation of Lp spaces (i.e., Hölder's inequality) for p<r<q is not sufficient to obtain (1.1). In fact, we would find that
max¯Ωf−min¯Ωf≤c‖∇f‖r,Ω≤c‖∇f‖αp,q,rp,Ω‖∇f‖1−αp,q,rq,Ω, | (1.2) |
where
αp,q,r=q−rr(q−p). |
Now, as r→N+, we see that αp,q,r tends to the exponent αp,q appearing in (1.1). However, in this limit, the first inequality in (1.2) fails to be true, as one can see by taking f(x)=loglog(1+1/|x|) in the unit ball B in RN for N≥2. In fact, f belongs to W1,N(B), but has infinite oscillation. Nevertheless, (1.1) still holds in the relevant case.
In order to prove (1.1), a different approach is needed. The one we present here has also the advantage to give a unified treatment for all the cases of p∈[1,∞]. These not only include the inequalities for 1≤p≤N and the noteworthy logarithmic profile in the threshold case p=N, but also the classical case p>N, which may be directly deduced from the classical Morrey-Sobolev embedding. In addition, our proof clearly unveils the dependence of the constant c in (1.1) on the relevant geometric parameters in hand.
Our proof holds when Ω is a bounded domain satisfying a uniform interior cone condition (see Section 2.2 for the definition). Notice that some regularity of Ω (or, alternatively, some information on the boundary traces) is needed for the validity of (1.1), as a simple counterexample shows. In fact, if in the planar domain
Ω={x=(x1,x2)∈R2:0<x1<1, |x2|<xr1},forr>1, |
we consider the function f(x)=x−1/q1, we see that f∈W1,q(Ω) if q<r. Nevertheless, the oscillation (and the L∞-norm) of f is infinite. Thus, for any q<∞, we can find r such that (1.1) fails on Ω.
The proof of (1.1) is given in Section 2 and is based on a pointwise bound on cones for f in terms of the Riesz potential of its gradient (see Lemma 2.1). When 1≤p≤N, (1.1) is obtained by combining that bound with an interpolation procedure performed on cones (see Lemma 2.5).
As an application of our inequalities, we shall use them to give an alternative way to obtain, and even improve, certain estimates proved by the authors in [10,Theorems 2.10 and 2.8]. These have been a crucial ingredient to obtain the stability of the spherical configuration for Alexandrov's Soap Bubble Theorem (SBT), Serrin's and other related overdetermined problems (see, e.g., [3,4,8,9,10,13,14,15]). More precisely, (1.1) can be used as a substitute of [15,Lemma 3.14] when aiming to obtain those stability results in the spirit of [10,Theorems 2.10 and 2.8]. We shall detail in Section 3 how this agenda can be carried out. We emphasize that, while [15,Lemma 3.14] can only be proved for sub-harmonic functions (see also [11] for a version for sub-solutions of elliptic equations in divergence form), our new bounds do not need this requirement. Thanks to this feature, they can also be useful in different and more general contexts. More on this will be clarified in forthcoming research. See also the recent paper [16].
In the remainder of this introduction, for the case of the SBT, we briefly describe the main steps of the argument that motivates the application of our interpolating inequalities. Alexandrov's SBT states that a closed surface Γ, embedded in RN, and that has constant mean curvature H must be a sphere. Roughly speaking, by stability of the spherical configuration in this problem, we mean an inequality of the type:
measureofclosenesstoasphere≤Ψ(‖H−H0‖). |
Here, Ψ is a non-negative continuous function vanishing at 0 and ‖H−H0‖ is the deviation of H from a reference constant H0, in a suitable norm. In the literature, there are many different ways to quantify the deviations of H from H0 and of a surface from a sphere (we refer the reader to the works [7,10,15] for a quite exhaustive list of references). It is clear that the weaker the norm ‖H−H0‖ is and the stronger the distance of Γ from a sphere is, the better the estimate is. On the other hand, in such a weak-strong setting, it may be difficult to obtain for Ψ the most desirable linear profile: Ψ(σ)=cσ. Here, c is some constant depending on some geometric parametes of the surface, easy to compute if possible. When this occurs, the optimality can be proved by considering sequences of ellipsoids.
In the works [8,9,10], we assume Γ to be the boundary of a bounded domain Ω, we set H0 to be the ratio |Γ|/N|Ω|, and we adopt an L2(Γ) (or even L1(Γ)) deviation of H from H0. Also, we measure the distance of Γ from a sphere, by the quantity ρe−ρi, where ρi and ρe, ρi≤ρe, are the radii of the best spherical annulus containing Γ. This will be given by Bρe(z)∖¯Bρi(z) for some z∈Ω. In other words, we obtained a bound of this type:
ρe−ρi≤cΨ(‖H−H0‖L2(Γ)). |
In this setting, in [10] we obtained a linear profile for Ψ in low dimension (N=2,3) and a Hölder profile with exponent 2/(N−2), for N≥5. For the threshold case N=4, we got, in a sense, a profile "arbitrarily close to a linear one" (see Remark 3.10 or [10], for details).
In Section 3 of this paper, for surfaces of class C2, we show that the interpolating bounds obtained in Section 2 help to improve the profile for N≥4. In fact, in Theorem 3.9, for N=4 we improve the older estimate (that was Ψ(σ)=cεσ1−ε, for any fixed ε>0) to a sharper and more plausible one: Ψ(σ)=cσlog(1/σ). Moreover, when N≥5, we are able to upgrade the profile Ψ(σ)=cσ2/(N−2) to Ψ(σ)=cεσ4/N−ε, for any fixed ε>0. This profile can be further improved to Ψ(σ)=cσ4/N, if we consider surfaces of class C2,γ, 1<γ≤1. For N=2,3, we just show that the new bounds in (1.1) provide an alternative way to recover the optimal profile previously obtained in [9,10].
Another novelty of this paper is that we show that our new improvements also hold if we enforce the quantity ρe−ρi by replacing it with the stronger deviation:
ρe−ρi+R‖ν−∇QzR‖2,Γ |
Here, R=1/H0, ν is the exterior unit normal vector to Γ, and Qz is defined by
Qz(x)=|x−z|22 for x,z∈RN. | (1.3) |
(Also in this case, the relevant norm is defined in the the corresponding normalized measure dSx/|Γ|.)
Thus, the smallness of this new measure of closeness to a sphere tells us not only that Γ is uniformly close to a sphere, but also that the Gauss map of Γ is quantitatively close in the average to that of the same sphere. Therefore, all in all, in Theorem 3.9, we enhance the last up-to-date bounds of [10] for the stability of the SBT as follows:
ρe−ρi+R‖ν−∇QzR‖2,Γ≤cΨ(‖H−H0‖L2(Γ)), | (1.4) |
where
Ψ(σ)={σ if N=2,3,σmax[log(1/σ),1] if N=4,στ if N≥5, | (1.5) |
where τ=4/N if Γ is of class C2,γ, 1<γ≤1. If Γ is of class C2, instead, when N≥5, we obtain that, for any sufficiently small ε>0, there exists a constant c=cε such that (1.4) and (1.5) hold with τ=4/N−ε. The constant c only depends on N, the diameter dΩ of Ω, and parameters associated with the assumed regularity of Γ. If Γ is of class C2, these are the radii ri and re of the uniform interior and exterior ball condition (see Section 3). If Γ is of class C2,γ, c depends on a suitable modulus of C2,γ-continuity for Γ. For details, see Theorem 3.9 and Remark 3.8. We stress that, for the second summand on the left-hand side of (1.4), we can actually obtain an optimal linear profile of stability, in every dimension (see (3.10)).
We spend a few final words to explain how the bounds derived in Section 2 come into play to obtain (1.4). To this aim, we let u∈C1(¯Ω)∩C2(Ω) be the solution of the problem:
Δu=N in Ω,u=0 on Γ. |
Also, we define the harmonic function h=u−Qz. Notice that, if z∈Ω, then we have that
12(|Ω||B|)1/N(ρe−ρi)≤12(ρ2e−ρ2i)=maxΓh−minΓh. |
Here, B is a unit ball in RN. Thus, a bound for the term ρe−ρi in (1.4) descends from the following identity
1N−1∫Ω|∇2h|2dx+1R∫Γ(uν−R)2dSx=∫Γ(H0−H)(uν)2dSx, | (1.6) |
which was proved in [8]. In fact, since the right-hand side can be easily bounded in terms of the L2(Γ) norm of H−H0, then the desired bound for ρe−ρi can be obtained if we can control the oscillation of h on Γ in terms of the first summand in (1.6). This goal is achieved by combining the bounds (1.1) (applied to h and its gradient) with some Poincaré-type inequality.
The second summand on the left-hand side of (1.4) can instead be estimated by observing that
|Rν−∇Qz|≤|Rν−uνν|+|∇u−∇Qz|=|R−uν|+|∇h| on Γ. |
The two quantities on the rightest-hand side can be estimated in L2(Γ)-norm by means of (1.6) and, again, by some inequalities derived in [10]. These involve a trace-type formula,
∫Γ|∇h|2dSx≤c∫Ω(−u)|∇2h|2dx, |
and another identity (stated in [8] and proved in [9]):
∫Ω(−u)|∇2h|2dx=12∫Γ(u2ν−R2)hνdSx. | (1.7) |
This last identity, immediately gives radial symmetry for Ω in Serrin's overdetermined problem (that prescribes that uν is constant on Γ, see [17,18]). Together with the arguments used to obtain (1.4), (1.7) will also help us to upgrade an analogous stability bound for radial symmetry in Serrin's problem. This task will be accomplished in Theorem 4.4.
Let SN−1 be the unit sphere in the Euclidean space RN, N≥2. For θ∈[0,π/2] and e∈SN−1, we set
Sθ={ω∈SN−1:cosθ<⟨ω,e⟩}. |
This is a spherical cap with axis e and opening width θ. We also denote by
Cx,a={x+aω:ω∈Sθ,0<s<a}, |
the finite right spherical cone with vertex at x, axis in some direction e, and height a>0. In what follows, |Cx,a| and |Sθ| will denote indifferently the N-dimensional Lebesgue measure of Cx,a and the (N−1)-dimensional surface measure of Sθ.
We start by proving some useful pointwise estimates in cones (see also [1]). In what follows, we set Cx=Cx,a and use the normalized Lebesgue measure dμy=dy/|E| for any measurable set E⊂RN of finite measure.
Lemma 2.1. For any f∈C1(¯Cx), it holds that
|f(x)−fCx|≤∫Cx|∇f(y)||y−x|N−1aN−|y−x|NNdμy. | (2.1) |
In particular, we have that
|f(x)−fCx|≤aNN∫Cx|∇f(y)||y−x|N−1dμy. | (2.2) |
Proof. By the change of variables y=x+sω for s∈(0,a) and ω∈Sθ, we write:
(x)−fCx=f(x)−∫Cxfdμy=1|Cx|∫Cx[f(x)−f(y)]dy=1|Cx|∫a0sN−1{∫Sθ[f(x)−f(x+sω)]dSω}ds, |
where dSω denotes the surface element on SN−1. Next, the fundamental theorem of calculus gives:
f(x)−f(x+sω)=−∫s0ω⋅∇f(x+tω)dt. |
Thus, we can infer that
|f(x)−fCx|≤1|Cx|∫a0∫SθsN−1[∫s0|∇f(x+tω)|dt]dSωds=1|Cx|∫a0sN−1∫Sθ[∫s0|∇f(x+tω)||x+tω−x|N−1tN−1dt]dSωds=1|Cx|∫a0sN−1[∫Cx,s|∇f(y)||y−x|N−1dy]ds. |
Now, by an application of Fubini's theorem we obtain that
∫a0(∫Cx,s|∇f(y)||y−x|N−1dy)ds=∫Cx|∇f(y)||y−x|N−1aN−|y−x|NNdy. |
Thus, (2.1) and (2.2) easily follow.
As a corollary, we have the following Morrey-Sobolev-type inequality. The relevant Lebesgue norms are defined with respect to the normalized measure dμy.
Corollary 2.2. If N<p≤∞ and f∈C1(¯Cx), we have that
|f(x)−fCx|≤aNβ(1−p′N′,p′+1)1/p′‖∇f‖p,Cx, | (2.3) |
where β(ξ,η) denotes Euler's beta function. When p=∞, (2.3) reads as
|f(x)−fCx|≤aNβ(1N,2)‖∇f‖∞,Cx, |
which can be obtained by taking the limit as p→∞ in (2.3).
Proof. The desired result follows from (2.1) by applying Hölder's inequality to the right-hand side and the calculation:
∫Cx(aN−|y−x|N|y−x|N−1)p′dμy=|Sθ||Cx|∫r0(aN−sNsN−1)p′sN−1ds=ap′∫10(1−t)p′t−N−1Np′dt=ap′β(1−N−1Np′,p′+1). |
Since β(ξ,η) is well-defined only if ξ,η>0, we get the restriction p>N. As already mentioned, the case p=∞ can be derived by taking the limit as p→∞.
Let Ω⊂RN be a bounded domain (i.e., a connected bounded open set) with boundary Γ. Given a>0 and θ∈[0,π/2], we say that Ω satisfies the (θ,a)-uniform interior cone condition, if for every x∈¯Ω there exists a cone Cx with opening width θ and height a, such that Cx⊂Ω and ¯Cx∩Γ={x}, whenever x∈Γ. The following result easily follows from Corollary 2.2.
Corollary 2.3. Let N<p≤∞ and Ω⊂RN be a bounded domain that satsfies the uniform interior (θ,a)-cone property.For every x∈¯Ω and f∈W1,p(Ω), we have that
|f(x)−fΩ|≤k(N,p,θ)a1−N/p|Ω|1/p‖∇f‖p,Ω, |
for some constant k(N,p,θ) only depending on N,p, and θ.
Proof. For any x∈¯Ω, there is a cone Cx contained in Ω. Hence, we apply (2.3) to the function f−fΩ+fCx and infer that
|f(x)−fΩ|≤aNβ(1−p′N′,p′+1)1/p′‖∇f‖p,Cx≤aNβ(1−p′N′,p′+1)1/p′(|Ω||Cx|)1/p‖∇f‖p,Ω≤β(1−p′N′,p′+1)1/p′N1/p′|Sθ|1/pa1−N/p|Ω|1/p‖∇f‖p,Ω. |
In the second inequality, we use the monotonicity of Lebesgue's integral with respect to set inclusion.
In this section, we aim to derive inequalities that bound from above the oscillation on ¯Ω of a function f with the Lp-norm of its gradient on Ω.
Theorem 2.4 (The case p>N). Set N<p≤∞. Let Ω⊂RN be a bounded domain satisfying the (θ,a)-uniform interior cone condition.
There exists a constant k(N,p,θ) only depending on N,p, and θ such that, for any f∈W1,p(Ω), it holds that
max¯Ωf−min¯Ωf≤k(N,p,θ)a1−N/p|Ω|1/p‖∇f‖p,Ω. | (2.4) |
Proof. Notice that the oscillation of f at the left-hand side of (2.4) is well defined, since f is continuous on ¯Ω.
Let xm,xM∈¯Ω be points at which f attains its minimum and maximum. Then, we have that
max¯Ωf−min¯Ωf≤f(xM)−fΩ+fΩ−f(xm) |
and we conclude by applying twice Corollary 2.3.
It is clear that the proof of Corollary 2.2 fails when 1≤p≤N, because of the singularity at x. However, in this case, we can still obtain a slightly different estimate by means of an interpolation procedure, if information on higher integrability of the gradient of f is available.
Lemma 2.5. Let f∈C1(¯Cx). Let 1≤p≤N, N<q≤∞, and set
αp,q=p(q−N)N(q−p). | (2.5) |
(i) If 1≤p<N, we have that
aN−1∫Cx|∇f(y)||y−x|N−1dμy≤kN,p,q‖∇f‖1−αp,qq,Cx‖∇f‖αp,qp,Cx, | (2.6) |
for some positive constant kN,p,q only depending on N,p, and q.
(ii) If p=N. we have that
aN−1 ∫Cx|∇f(y)||y−x|N−1dμy≤2qq−N‖∇f‖N,Cxlog(e‖∇f‖q,Cx‖∇f‖N,Cx). | (2.7) |
Proof. For any σ∈(0,a), we compute that
∫Cx|∇f(y)||y−x|N−1dy=∫Cx,σ|∇f(y)||y−x|N−1dy+∫Cx∖Cx,σ|∇f(y)||y−x|N−1dy≤[∫Cx,σdy|y−x|q′(N−1)]1/q′(∫Cx,σ|∇f(y)|qdy)1/q+[∫Cx∖Cx,σdy|y−x|p′(N−1)]1/p′(∫Cx∖Cx,σ|∇f|pdy)1/p, | (2.8) |
by Hölder's inequality. Now, a direct computation shows that
[∫Cx,σdy|y−x|q′(N−1)]1/q′=[q−1q−N|Sθ|]1/q′σq−Nq,[∫Cx∖Cx,σdy|y−x|p′(N−1)]1/p′={[p−1N−p|Sθ|(σ−N−pp−1−a−N−pp−1)]1/p′if 1≤p<N,[|Sθ|logaσ]1/N′if p=N. | (2.9) |
For p=1, this formula must be intended in the limit as p→1.
(ⅰ) Let 1≤p<N. By (2.8), (2.9), and some algebraic manipulations, we can infer that
aN−1∫Cx|∇f(y)||y−x|N−1dμy≤[N(q−1)q−N]1−1/q‖∇f‖q,Cx(σa)1−N/q+[N(p−1)N−p]1−1/p‖∇f‖p,Cx(σa)1−N/p | (2.10) |
for any σ∈(0,a]. The minimum of the right-hand side is attained either at
¯σ=a[N(p−1)N−p]q(p−1)N(q−p)[q−NN(q−1)]p(q−1)N(q−p)(1−αp,qαp,q‖∇f‖p,Cx‖∇f‖q,Cx)pqN(q−p), |
or at σ=a. In the former case, we plug ¯σ into (2.10) and obtain (2.6) with some computable constant k′. In the latter case, we have that
[N(p−1)N−p]q(p−1)N(q−p)[q−NN(q−1)]p(q−1)N(q−p)(1−αp,qαp,q‖∇f‖p,Cx‖∇f‖q,Cx)pqN(q−p)>1, |
since ¯σ>a. Hence, by means of this inequality and the fact that we have that
∫Cx|∇f(y)||y−x|N−1dμy≤1aN−1[N(q−1)q−N]1−1/q‖∇f‖q,Cx(σa)1−N/q, |
thanks to (2.8), we again obtain (2.6) for some possibly different computable constant k″. Thus, we conclude that (2.6) holds true with kN,p,q=max(k′,k″).
(ⅱ) Let p=N. We proceed as in the case (ⅰ) by putting together (2.8) and (2.9). After some calculation, we obtain:
aN−1∫Cx|∇f(y)||y−x|N−1dμy≤[N(q−1)q−N]1−1/q‖∇f‖q,Cx(σa)1−N/q+(Nlogaσ)1−1/N‖∇f‖N,Cx. |
If we assume that 0<σ<a/e, being as N(q−1)≥q−N, we can simplify this inequality to get that
aN−1N∫Cx|∇f(y)||y−x|N−1dμy≤q−1q−N‖∇f‖q,Cx(σa)1−N/q+ ‖∇f‖N,Cxlogaσ. | (2.11) |
Thus, the minimum of the right-hand side is attained either at
σ=¯σ=a[q′‖∇f‖N,Cx‖∇f‖q,Cx]qq−N orat σ=a/e. |
In the former case, we get that
aN−1N∫Cx|∇f(y)||y−x|N−1dμy≤2qq−N‖∇f‖N,Cxlog(e‖∇f‖q,Cxq′‖∇f‖N,Cx), |
and hence (2.7) holds true, being as q′≥1. In the latter case, we have that
e−1≤[q′‖∇f‖N,Cx‖∇f‖q,Cx]qq−N, |
since ¯σ≥a/e. Thus, we get that
aN−1N∫Cx|∇f(y)||y−x|N−1dμy≤q−1q−N‖∇f‖q,CxeN/q−1+‖∇f‖N,Cx≤2q−Nq−N‖∇f‖N,Cx≤2q−Nq−N‖∇f‖N,Cxlog(e‖∇f‖q,Cx‖∇f‖N,Cx), |
being as ‖∇f‖N,Cx≤‖∇f‖q,Cx. Since 2q−N≤2q, (2.7) still holds true.
We obtain the following consequence.
Corollary 2.6. For any cone Cx⊂Ω of height a and opening width θ, it holds that
∫Cx|∇f(y)||y−x|N−1dμy≤k(N,p,q,θ)|Ω|a2N−1‖∇f‖1−αp,qq,Ω‖∇f‖αp,qp,Ω, |
for 1≤p<N and, if p=N,
∫Cx|∇f(y)||y−x|N−1dμy≤k(N,p,q,θ)|Ω|1/NaN‖∇f‖N,Ωlog(e‖∇f‖q,Ω‖∇f‖N,Ω), |
for some constant k(N,p,q,θ) only depending on N,p,q, and θ.
Proof. The monotonicity of Lebesgue's measure with respect to set inclusion and (2.11) easily give:
aN−1N∫Cx|∇f(y)||y−x|N−1dμy≤q−1q−N(|Ω||Cx|)1/q‖∇f‖q,Ω(σa)1−N/q+(|Ω||Cx|)1/N ‖∇f‖N,Ωlogaσ. |
By using that (|Ω|/|Cx|)1/q−1/N≤1 (being as Cx⊂Ω and q>N), the above inequality becomes:
aN−1N∫Cx|∇f(y)||y−x|N−1dμy≤(|Ω||Cx|)1/N{q−1q−N‖∇f‖q,Ω(σa)1−N/q+‖∇f‖N,Ωlogaσ}. |
Thus, we can proceed as in the last part of the proof of Lemma 2.5, with similar algebraic manipulations.
In light of Corollary 2.6, we can somewhat extend the bound (2.4) to the case 1≤p≤N, provided f∈W1,q(Ω) for q>N. The proof is straightforward and runs as that of Theorem 2.4.
Theorem 2.7. Let 1≤p≤N, N<q≤∞, and set αp,q as in (2.5). Let Ω⊂RN be a bounded domain satisfying the (θ,a)-uniform interior cone condition.
For any f∈W1,q(Ω), it holds that
max¯Ωf−min¯Ωf≤k(N,p,q,θ)|Ω|1/paN/p−1‖∇f‖αp,qp,Ω‖∇f‖1−αp,qq,Ω, |
if 1≤p<N and, if p=N,
max¯Ωf−min¯Ωf≤k(N,p,q,θ)|Ω|1/NaN‖∇f‖N,Ωlog(e‖∇f‖q,Ω‖∇f‖N,Ω). |
Here, k(N,p,q,θ) is some constant only depending on N,p,q, θ.
As already mentioned, Theorems 2.4 and 2.7 give an alternative way to obtain, and even upgrade, the bounds in [10,Theorems 2.10 and 2.8]. As a by-product, we also obtain new upgraded versions of stability estimates for the Soap Bubble Theorem and Serrin's symmetry result. In this and the next section, we shall give some details on how to obtain the new versions of those stability results. Of course, a similar reasoning can be applied to other stability results contained in [4,8,10,15].
For a point z∈Ω, ρi and ρe shall denote the radius of the largest ball contained in Ω and that of the smallest ball that contains Ω, both centered at z; in formulas,
ρi=minx∈Γ|x−z| and ρe=maxx∈Γ|x−z|. | (3.1) |
We say that Ω satisfies a uniform interior sphere condition (with radius r) if for every p∈Γ there exists a ball Br⊂Ω such that ∂Br∩Γ={p}; Ω satisfies a uniform exterior sphere condition if RN∖¯Ω satisfies a uniform interior sphere condition. From now on, we will consider a bounded domain Ω with boundary Γ of class C2, so that Ω satisfies both a uniform interior and exterior sphere condition. We shall denote by ri and re the relevant respective radii. It is trivial to check that when Ω satisfies the interior condition with radius ri, then it satisfies the uniform interior (θ,a)-cone condition with
θ=√22,a=ri. | (3.2) |
Next, we consider the solution u∈C0(¯Ω)∩C2(Ω) of
Δu=N in Ω,u=0 on Γ. | (3.3) |
It is well-known that u∈Cm,γ(¯Ω) if Γ is of class Cm,γ, 0<γ≤1, for m=1,2,⋯.
By M we denote a uniform upper bound for the gradient of u on ¯Ω, in formulas,
M≥max¯Ω|∇u|=maxΓuν. |
As shown in [8,Theorem 3.10], we can choose an explicit value for M:
M=(N+1)dΩ(dΩ+re)2re. | (3.4) |
By following [10,15], we consider the harmonic function
h=Qz−u, |
where Qz is defined in (1.3). Notice that, if z∈Ω, it holds that
maxΓh−minΓh=12(ρ2e−ρ2i)≥(|Ω||B|)1/Nρe−ρi2≥ri2(ρe−ρi). | (3.5) |
The left-hand side of this inequality can be estimated by Theorems 2.4 and 2.7.
As in [10], it will be convenient to choose z∈Ω as a global minimum point of u. We know from [12] that, in this case, the distance δΓ(z) of z to Γ can be estimated from below in terms of the inradius rΩ (the radius of a maximal ball contained in Ω). In fact, in light of [12,Theorem 1.1], it holds that
δΓ(z)≥rΩ√N, | (3.6) |
if Γ is mean convex (i.e., H≥0). If Γ is a general surface of class C2, [12,Corollary 2.7] gives instead the slightly poorer bound:
δΓ(z)≥rΩ√N[1+N2−12NdΩre(1+dΩre)]−1/2. | (3.7) |
Remark 3.1 (On the normalized norms). For the sake of consistency with the previous sections, we will continue to denote by ‖⋅‖p,Ω and ‖⋅‖p,Γ the Lp-norms in the relevant normalized measure. Since it holds that
|B|rNΩ≤|Ω|≤|B|dNΩ and N|B|rN−1Ω≤|Γ|≤N|Ω|ri, |
such norms are equivalent to the standard ones. The first three inequalities follow from the inclusions BrΩ⊂Ω⊂BdΩ. The last inequality is obtained by putting together the identity
N|Ω|=∫ΓuνdSx |
with the inequality uν≥ri, which holds true at any point in Γ, by an adaptation of Hopf's lemma (see [8,Theorem 3.10]).
Notice that, since rΩ≥ri, rΩ can be replaced by ri in all the relevant formulas.
In what follows, we use the letter c to denote a constant whose value may change line by line. The dependence of c on the relevant parameters will be indicated whenever it is important. All the constants c can be explicitly computed (by following the steps in the relevant proofs) and estimated in terms of the indicated parameters only.
By applying Theorems 2.4 and 2.7 to h, we easily obtain the starting point of our analysis.
Lemma 3.2. Let Ω⊂RN, N≥2, be a bounded domain with boundary Γ of class C2. Letz be a point in Ω, andconsider the function h=Qz−u, with Qz defined in (1.3).
There exists a constant c=c(N,p,ri) such that
ρe−ρi≤c{‖∇h‖p,Ω ifp>N;‖∇h‖N,Ωlog(e‖∇h‖∞,Ω‖∇h‖N,Ω) ifp=N;‖∇h‖(N−p)/N∞,Ω‖∇h‖p/Np,Ω if1≤p<N. |
Proof. We apply Theorems 2.4 and 2.7, with f=h and q=∞. By taking into account (3.5) and (3.2), the desired estimates easily follow. (Notice that (3.2) informs us that in Theorems 2.4 and 2.7 we can take a=ri.)
Remark 3.3 (Weighted Poincaré inequality). Here, we recall a bound for the gradient of h, which we will need in the sequel. Since z is a critical point of h (being as ∇h(z)=∇Qz(z)−∇u(z)=0), we know from [10,Corollary 2.3] that h satisfies the weighted Poincaré inequality
‖∇h‖r,Ω≤c‖δαΓ∇2h‖p,Ω. |
Here, r,p,α are three numbers such that
1≤p≤r≤NpN−p(1−α),p(1−α)<N,0≤α≤1. |
The constant c can be explicitly estimated by putting together item (ⅲ) of [10,Remark 2.4], (3.7), and the normalizations discussed in Remark 3.1. In detail, we can compute that
c≤kN,r,p,α|Ω|1−αN(dΩ/ri)N[N+(N2−1)dΩ2re(1+dΩre)]N/2, |
for some constant kN,r,p,α only depending on N,r,p,α. When Γ is mean convex, by using (3.6) in place of (3.7), we obtain the finer bound:
c≤kN,r,p,α|Ω|1−αN(dΩ/ri)N. |
As described in the introduction, in order to obtain stability estimates for the Soap Bubble Theorem, we must associate the difference ρe−ρi with the L2-norm of the hessian matrix ∇2h. The following result gives this association.
Theorem 3.4. Let Ω⊂RN, N≥2, be a bounded domain with boundary Γ of class C2. Let z∈Ω be a global minimum point of u in ¯Ω and set h=Qz−u.Then, there exists a constant c=c(N,ri,re,dΩ) such that
ρe−ρi≤c{‖∇2h‖2,Ω for N=2,3;‖∇2h‖2,Ωmax[log(e‖∇h‖∞,Ω‖∇2h‖2,Ω),1], for N=4;‖∇h‖N−4N−2∞,Ω‖∇2h‖2N−22,Ω, for N≥5. |
Proof. (ⅰ) Lemma 3.2 with p=6 gives that
ρe−ρi≤c‖∇h‖6,Ω≤c‖∇2h‖2,Ω. |
The last inequality follows from Remark 3.3 with r=6, p=3/2, and α=0, and Hölder's inequality, for N=2, and directly from Remark 3.3 with r=6, p=2, and α=0, for N=3.
(ⅱ) Let N=4. We use Lemma 3.2 with p=N=4 and get:
ρe−ρi≤cmax{‖∇2h‖4,Ωlog(e‖∇h‖∞,Ω‖∇2h‖4,Ω),‖∇2h‖4,Ω}. |
Next, Remark 3.3 with r=4, p=2, α=0, gives:
‖∇h‖4,Ω≤c‖∇2h‖2,Ω. |
Thus, the desired conclusion ensues by invoking the monotonicity of the function t↦tmax{log(A/t),1} for every A>0.
(ⅲ) When N≥5, we can use Lemma 3.2 with p=2N/(N−2) and put it together with Remark 3.3 with r=2N/(N−2), p=2, and α=0.
Remark 3.5. For N≥4 the estimates of this theorem depend on ‖∇h‖∞,Ω. Thus, as done in [10], since we know that
‖∇h‖∞,Ω≤M+dΩ, |
we can easily bound ρe−ρi in terms of some constant (which possibly depends on ri,re, and dΩ, thanks to (3.4)) and the number ‖∇2h‖2,Ω. Thanks to identity (1.6), this number is connected to the deviation H−H0. This will lead to the asymptotic profile in the quantitative symmetry estimate for the Soap Bubble Theorem obtained in [10], with an improvement for the case N=4.
However, notice that, when Ω is near a ball in some good topology, the function h tends to be a constant, and hence ‖∇h‖∞,Ω tends to be zero. Thus, we expect to improve the relevant bounds in Theorem 3.4, once we can control ‖∇h‖∞,Ω in terms of ‖∇2h‖2,Ω. This control will in turn benefit the quantitative symmetry estimate we are aiming to. It turns out that an adaptation of our Theorem 2.7 gives such desired bound for ‖∇h‖∞,Ω, if an a priori bound for ‖∇2h‖q,Ω for large q is available, as the following corollary states.
Corollary 3.6. Let Ω⊂RN be a bounded domain with boundary of class C2. Let 1≤p<N, N<q≤∞, and set αp,q as in (2.5). Then, if h∈W2,q(Ω), it holds that
‖∇h‖∞,Ω≤c|Ω|1/prN/p−1i‖∇2h‖αp,qp,Ω‖∇2h‖1−αp,qq,Ω. |
Here, c is a constant only depending on N, p, q.
Proof. Since Γ is of class C2, Ω has the uniform interior cone property with θ=√2/2 and a=ri. Let x∈¯Ω and let ℓ be any unit vector. Applying Theorem 2.7 and using that, with our choice of z, |hℓ(x)|=|hℓ(x)−hℓ(z)|, we have that
|hℓ(x)|≤k(N,p,q)|Ω|1/prN/p−1i‖∇hℓ‖αp,qp,Ω‖∇hℓ‖1−αp,qq,Ω≤k(N,p,q)|Ω|1/prN/p−1i‖∇2h‖αp,qp,Ω‖∇2h‖1−αp,qq,Ω, |
where we used the pointwise inequality |∇hℓ|≤|∇2h|. Hence, taking the supremum over all directions ℓ yields the desired conclusion.
An inspection of the proof tells us that the corollary could be stated for a domain satisfying an interior cone condition.
This corollary allows us to upgrade Theorem 3.4 for N≥5. Notice that, for N=4, we would not get any subtantial improvement, due to the presence of the logarithm in the relevant claim of that theorem.
Corollary 3.7. Let Ω⊂RN, N≥5, be a bounded domain with boundary Γ of class C2. Let z∈Ω be a global minimum point of u in ¯Ω, set h=Qz−u, and suppose that h∈W2,q(Ω).Then, for every q∈(N,∞], there exists a constant c=c(N,q,ri,re,dΩ) such that
ρe−ρi≤c‖∇2h‖q(N−4)(q−2)Nq,Ω‖∇2h‖4N−2(N−4)N(q−2)2,Ω. |
Proof. Our claim simply follows by combining Theorem 3.4 and Corollary 3.6 with the choice p=2.
We are now in position to obtain our new quantitative estimates of radial symmetry per the Soap Bubble Theorem. As already mentioned, all we have to do is to relate the norm ‖∇2h‖2,Ω to the deviation of H from H0 in some norm.
The quantities ‖∇h‖∞,Ω and ‖∇2h‖q,Ω in Theorem 3.4 and Corollary 3.7 will contribute to the computation of the constant in the desired stability profile, as explained in the next remark.
Remark 3.8. We shall consider two regularity assumptions on Γ.
(ⅰ) When Γ is of class C2, we have that u∈W2,q(Ω) for any q∈[1,∞) and an a priori bound for ‖∇2h‖q,Ω can be obtained, by the standard Lq estimates for elliptic equations, being as ∇2h=I−∇2u. In fact, by putting together [6,Theorems 914 and 9.15], even under the weaker assumption of Γ∈C1,1, we can obtain for u the bound
‖∇2u‖q,Ω≤C for N<q<∞, |
where C only depends on N,q, |Ω|, and the regularity Ω (and may blow up as q→∞). It is well known that Γ is of class C1,1 if and only if it satisfies both the interior and exterior ball condition. Thus, we can claim that C only depends on N,q,dΩ,ri, and re.
(ⅱ) When Γ is of class C2,γ with 0<γ≤1, we can obtain an a priori bound also for ‖∇2h‖∞,Ω, by standard Schauder's estimates for ∇2u (see [6]), in terms of the C2,γ-modulus of continuity ω2,γ of Γ. (For a definition of ω2,γ, see e.g., [2,Remark 1].)
The following theorem clearly gives (1.4).
Theorem 3.9 (Soap Bubble Theorem: enhanced stability). Let N≥2 and let Ω⊂RN be a bounded domain with boundary Γ of class C2. Denote by H the mean curvature of Γ and set R=N|Ω|/|Γ| and H0=1/R.
Let z∈Ω be a global minimum point of the solution u of (3.3) and let ρi and ρe be defined by (3.1). Then, the following inequalities hold true.
(i) If 2≤N≤4, there exists a constant c=c(N,dΩ,ri,re) such that
ρe−ρi≤c{‖H0−H‖2,Γ, if N=2,3,‖H0−H‖2,Γmax[log(1‖H0−H‖2,Γ),1], if N=4. | (3.8) |
(ii) If N≥5, for any q∈(N,∞), there exists a constant c=c(N,q,dΩ,ri,re) such that
ρe−ρi≤c‖H0−H‖4N−2(N−4)N(q−2)2,Γ. | (3.9) |
Moreover, (for any N≥2) we have that
R‖ν−∇QzR‖2,Γ≤c‖H0−H‖2,Γ. | (3.10) |
If Γ is of class C2,γ, 0<γ≤1, the exponent in (3.9) can be replaced by its limit as q→∞, i.e., 4/N.In this case, the relevant constant c only depends on N, dΩ, and the C2,γ-modulus of continuity of Γ.
Proof. Inequalities (3.8) and (3.9) will simply follow from the inequality:
‖∇2h‖2,Ω≤c‖H−H0‖2,Γ. | (3.11) |
This was proved in [10].
For the reader's convenience, we summarize the main steps in the proof of [10,Theorem 3.5], which lead to (3.11), with the necessary modifications. As usual, the constant c may change from line to line and only depends on quantities (e.g., R, ‖uν‖∞,Γ, ‖Qzν‖∞,Γ) that, in turn, can be bounded in terms of the parameters indicated in the statement.
The starting point is a modification of the fundamental identity (1.6):
1N−1∫Ω|∇2h|2dx+1R∫Γ(uν−R)2dSx=−∫Γ(H0−H)hνuνdSx+∫Γ(H0−H)(uν−R)QzνdSx. |
Next, if we discard the first summand in this identity, by Cauchy-Schwarz inequality we obtain that
‖uν−R‖22,Γ≤c‖H−H0‖2,Γ(‖hν‖2,Γ+‖uν−R‖2,Γ). | (3.12) |
Instead, if we discard the second summand, we can infer that
∫Ω|∇2h|2dx≤c‖H−H0‖2,Γ(‖hν‖2,Γ+‖uν−R‖2,Γ). | (3.13) |
Now, we use the fact that we can control ∇h (and hence hν) on Γ in terms of the deviation uν−R. This is obtained by combining a trace-type inequality for h derived in [10,Lemma 2.5] and identity (1.7), as follows:
∫Γ|∇h|2dSx≤c∫Ω(−u)|∇2h|2dx=12c∫Γ(u2ν−R2)hνdSx≤c‖uν−R‖2,Γ‖hν‖2,Γ≤c‖uν−R‖2,Γ‖∇h‖2,Γ. |
This then gives:
‖hν‖2,Γ≤‖∇h‖2,Γ≤c‖uν−R‖2,Γ. | (3.14) |
Thus, inserting this inequality into (3.12) gives that
‖uν−R‖2,Γ≤c‖H−H0‖2,Γ. | (3.15) |
Also, by plugging it into (3.13), we infer that
∫Ω|∇2h|2dx≤c‖H−H0‖2,Γ‖uν−R‖2,Γ≤c‖H−H0‖22,Γ. |
Therefore, (3.11) follows at once.
Now, we proceed to prove (3.8) and (3.9). The cases N=2,3 easily follow from Theorem 3.4. Thus, we are left to prove it for N≥4.
For N=4, we simply combine Theorem 3.4 and the first part of Remark 3.5. Indeed, ‖∇h‖∞,Ω is bounded by a constant which only depends on ri,re, and dΩ.
For N≥5, instead, we use Corollary 3.7 and Remark 3.8, which give
ρe−ρi≤c‖∇2h‖4N−2(N−4)N(q−2)2,Ω. |
Hence, (3.9) ensues from (3.11). The case in which Γ is of class C2,γ can be dealt similarly.
To conclude the proof, we are left to show that (3.10) also holds. To this aim, as done in the introduction, we observe that
|ν(x)−x−zR|≤|R−uν(x)|+|∇h(x)|R for x∈Γ. |
Hence, we infer that
R(∫Γ|ν(x)−x−zR|2dSx|Γ|)1/2≤‖uν−R‖2,Γ+‖∇h‖2,Γ≤c‖uν−R‖2,Γ, |
where we applied the triangle inequality and the second inequality in (3.14). By using (3.15), then (3.10) easily follows from the last inequality above.
Remark 3.10. In order to compare the results of Theorem 3.9 to previous estimates, we recall what we obtained in [10,Theorem 3.5] — the last up-to-date bound for stability in the Soap Bubble Theorem. In fact, there we obtained the bound
ρe−ρi≤cΨ(‖H−H0‖L2(Γ)), |
with
Ψ(σ)={σ if N=2,3,σ1−ε if N=4,σ2/(N−2) if N≥5, |
where the case N=4 must be interpreted thus: for any 0<ε<1, there exists a constant c=cε (which may blow up as ε→0), such that case N=4 holds. Theorem 3.9 clearly improves these profiles if Γ is either of class C2 or C2,γ. Moreover, it also states that we can control linearly the deviation of the Gauss map from that of a sphere, at least in the L2-norm.
In order to obtain stability estimates for Serrin's problem, we must use identity (1.7). In fact, this relates the weighted integral at the right-hand side to the deviation uν−R. Since the torsion u can be easily bounded below by δΓ (see [9,Lemma 3.1]), we understand that this time we must associate the difference ρe−ρi with the weighted L2-norm ‖δ1/2Γ∇2h‖2,Ω. The following result goes in that direction.
Theorem 4.1. Let Ω⊂RN, N≥2, be a bounded domain with boundary Γ of class C2 and z∈Ω be a global minimum point of the solution u of (3.3). Consider the function h=Qz−u, with Qz given by (1.3). Then, there exists a constant c=c(N,dΩ,ri,re) such that
ρe−ρi≤c{‖δ1/2Γ∇2h‖2,Ω if N=2;‖δ1/2Γ∇2h‖2,Ωmax[log(e‖∇h‖∞,Ω‖δ1/2Γ∇2h‖2,Ω),1] if N=3;‖∇h‖(N−3)/(N−1)∞,Ω‖δ1/2Γ∇2h‖2/(N−1)2,Ω if N≥4. |
Proof. (ⅰ) Let N=2. By using Lemma 3.2 with p=4 we have that
ρe−ρi≤c‖∇h‖4,Ω. |
By applying Remark 3.3 with r=4, p=2, and α=1/2, we obtain that
‖∇h‖4,Ω≤c‖δ1/2Γ∇2h‖2,Ω, |
and the conclusion follows.
(ⅱ) Let N=3. By using Remark 3.3 with r=3, p=2, α=1/2, we get
‖∇h‖3,Ω≤c‖δ1/2Γ∇2h‖2,Ω. |
The conclusion follows by using Lemma 3.2 with p=N=3.
(ⅲ) When N≥4, we use Lemma 3.2 with p=2N/(N−1) and put it together with Remark 3.3 with r=2NN−1, p=2, α=1/2.
By recalling Remark 3.5, to gain better stability for Serrin's problem for N≥3, we need to obtain a bound similar to that in Corollary 3.6, but with ‖∇2h‖p,Ω replaced by ‖δ1/2Γ∇2h‖p,Ω. This time, we proceed differently.
Lemma 4.2. Set 1≤p≤∞ and q>N. Let Ω⊂RN, N≥2, be a bounded domain with boundary Γ of class C2 and assume that h∈W2,q(Ω).Then, there exists a constant c=c(N,p,q) such that
‖∇h‖N+p(1−N/q)∞,Ω≤c|Ω|‖∇h‖p(1−N/q)p,Ω‖∇2h‖Nq,Ω. | (4.1) |
Proof. For any x∈¯Ω there is a cone Cx,a⊂Ω. Applying (2.3) with p=q to any cone Cx,σ⊂Cx,a gives that
|f(x)|≤∫Cx,σ|f|dμy+cσ‖∇f‖q,Cx,σ≤‖f‖p,Cx,σ+cσ‖∇f‖q,Cx,σ, |
where we used Hölder's inequality at the second inequality. Here, c=c(N,q). Thus, we have that
max¯Ωf−min¯Ωf≤2max¯Ω|f|≤c(|Ω|1/pσ−N/p‖f‖p,Ω+c|Ω|1/qσ1−N/q‖∇f‖q,Ω), |
for every σ∈(0,a), where in the second inequality we also used the monotonicity of Lebesgue's integral with respect to set inclusion. Here, c=c(N,p,q) (notice that the dependence on θ can be dropped, since θ=√2/2, being as Γ of class C2). We now minimize in σ as done before. This time, we omit the details. We end up with the formula:
max¯Ωf−min¯Ωf≤c|Ω|1N+p(1−N/q)‖f‖p(1−N/q)N+p(1−N/q)p,Ω‖∇f‖NN+p(1−N/q)q,Ω. |
This holds for any x∈¯Ω and 1≤p<q≤∞. By choosing f as any directional derivative hℓ of h and using that, with our choice of z, |hℓ(x)|=|hℓ(x)−hℓ(z)|, we thus get that
|hℓ(x)|N+p(1−N/q)≤c|Ω|‖hℓ‖p(1−N/q)p,Ω‖∇hℓ‖Nq,Ω forany x∈¯Ω. |
Hence, (4.1) follows by observing that |hℓ|≤|∇h|, |∇hℓ|≤|∇2h|, and by choosing ℓ such that hℓ(x)=|∇h(x)| and x∈Γ that maximizes |∇h| on ¯Ω.
As for Corollary 3.6, the lemma could be stated for a domain satisfying an interior cone condition.
Corollary 4.3. Set 1≤p<2N and q>N. Under the assumptions of Lemma 4.2, we have that
‖∇h‖2N−p+2p(1−N/q)∞,Ω≤c‖∇2h‖2N−pq,Ω‖δ1/2Γ∇2h‖2p(1−N/q)p,Ω. | (4.2) |
Here, the constant c only depends on N, p, q, dΩ, ri, and re.
Proof. We use Remark 3.3 with r, p, and α replaced by 2pN/(2N−p), p, and 1/2, respectively. We thus get that
‖∇h‖2pN2N−p,Ω≤c‖δ1/2Γ∇2h‖p,Ω. |
Therefore, (4.2) follows by combining this bound and (4.1) with p replaced by 2pN/(2N−p).
Theorem 4.4 (Serrin's problem: enhanced stability). Let Ω⊂RN, N≥2, be a bounded domain with boundary Γ of class C2 and setR=N|Ω|/|Γ|.
Let u be the solution of problem (3.3) and z∈Ω be a global minimum point of u in ¯Ω.Then, there exists a constant c=c(N,dΩ,ri,re) such that
ρe−ρi≤c{‖uν−R‖2,Γ if N=2;‖uν−R‖2,Γmax[log(1‖uν−R‖2,Γ),1] if N=3. |
When N≥4, for any q∈(N,∞), there exists a constant c=c(N,q,dΩ,ri,re) such that
ρe−ρi≤c‖uν−R‖4−2N/qN+1−2N/q2,Γ. | (4.3) |
Moreover (for any N≥2),
R‖ν−∇QzR‖2,Γ≤c‖uν−R‖2,Γ, |
for some constant c=c(N,dΩ,ri,re).
If Γ is of class C2,γ, 0<γ≤1, the stability exponent in (4.3) for N≥4 can be replaced its limit as q→∞, i.e., 4/(N+1).In this case, c only depends on N, dΩ, and the C2,γ-modulus of continuity of Γ.
Proof. It is sufficient to notice that, thanks to (1.7) and the pointwise inequality δΓ≤−2u/ri, we can infer that
‖δ1/2Γ∇2h‖22,Ω≤c∫Ω(−u)|∇2h|2dx≤c‖uν−R‖2,Γ‖hν‖2,Γ. |
Thus, by (3.14), we obtain that
‖δ1/2Γ∇2h‖2,Ω≤c‖uν−R‖2,Γ. |
Therefore, with this inequality in hand, we can proceed similarly to the proof of Theorem 3.9 by also taking into account Remark 3.8. For instance, the claim for N≥4 simply follows from Theorem 4.1 and Corollary 4.3 with p=2.
The remaining claims follow from Theorem 4.1 at once.
Remark 4.5. In order to compare the results of Theorem 4.4 to previous estimates, it is sufficient to recall what we obtained in [10,Theorem 3.1] — the last up-to-date bound for stability in Serrin's problem. In fact, there we obtained the bound
ρe−ρi≤cΨ(‖uν−R‖L2(Γ)), |
with
Ψ(σ)={σ if N=2,σ1−ε if N=3,σ2/(N−1) if N≥4. |
The case N=3 must be interpreted thus: for any 0<ε<1 there exists a constant c=cε (which may blow up as ε→0), such that case N=3 holds.
The comparison with Theorem 4.4 is left to the reader.
As already mentioned in the introduction for the Soap Bubble Theorem, if one adopts a stronger norm for the deviation uν−R, linear stability can also be obtained (with some restrictions) in general dimension. See for instance [5].
Remark 4.6. A direct inspection of the corresponding proofs tells us that the dependence of the relevant constant c on the parameter re can be removed whenever Γ is mean convex. In fact, in this case, the bounds in (3.4), (3.7) and the former inequality for c in Remark 3.3 can be replaced by [10,Formula (2.4)], (3.6) and the latter inequality for c in Remark 3.3. $
Rolando Magnanini was partially supported by the Gruppo Nazionale per l'Analisi Matematica, la Probabilità e le loro Applicazioni (GNAMPA) of the italian Istituto Nazionale di Alta Matematica (INdAM). Giorgio Poggesi is supported by the Australian Laureate Fellowship FL190100081 "Minimal surfaces, free boundaries and partial differential equations" and is member of AustMS and INdAM/GNAMPA.
The authors declare no conflict of interest.
[1] | R. A. Adams, Sobolev spaces, New York: Academic Press, 1975. |
[2] | A. Aftalion, J. Busca, W. Reichel, Approximate radial symmetry for overdetermined boundary value problems, Adv. Differential Equations, 4 (1999), 907–932. |
[3] | L. Cavallina, G. Poggesi, T. Yachimura, Quantitative stability estimates for a two-phase Serrin-type overdetermined problem, 2021, arXiv: 2107.05889. |
[4] |
S. Dipierro, G. Poggesi, E. Valdinoci, A Serrin-type problem with partial knowledge of the domain, Nonlinear Anal., 208 (2021), 112330. http://dx.doi.org/10.1016/j.na.2021.112330 doi: 10.1016/j.na.2021.112330
![]() |
[5] |
A. Gilsbach, M. Onodera, Linear stability estimates for Serrin's problem via a modified implicit function theorem, Calc. Var., 60 (2021), 241. http://dx.doi.org/10.1007/s00526-021-02107-1 doi: 10.1007/s00526-021-02107-1
![]() |
[6] | D. Gilbarg, N. S. Trudinger, Elliptic partial differential equations of second order, Berlin, Heidelberg: Springer, 2001. http://dx.doi.org/10.1007/978-3-642-61798-0 |
[7] |
R. Magnanini, Alexandrov, Serrin, Weinberger, Reilly: symmetry and stability by integral identities, Bruno Pini Mathematical Analysis Seminar, 8 (2017), 121–141. http://dx.doi.org/10.6092/issn.2240-2829/7800 doi: 10.6092/issn.2240-2829/7800
![]() |
[8] |
R. Magnanini, G. Poggesi, On the stability for Alexandrov's Soap Bubble theorem, JAMA, 139 (2019), 179–205. http://dx.doi.org/10.1007/s11854-019-0058-y doi: 10.1007/s11854-019-0058-y
![]() |
[9] |
R. Magnanini, G. Poggesi, Serrin's problem and Alexandrov's Soap Bubble Theorem: enhanced stability via integral identities, Indiana Univ. Math. J., 69 (2020), 1181–1205. http://dx.doi.org/10.1512/iumj.2020.69.7925 doi: 10.1512/iumj.2020.69.7925
![]() |
[10] |
R. Magnanini, G. Poggesi, Nearly optimal stability for Serrin's problem and the Soap Bubble theorem, Calc. Var., 59 (2020), 35. http://dx.doi.org/10.1007/s00526-019-1689-7 doi: 10.1007/s00526-019-1689-7
![]() |
[11] | R. Magnanini, G. Poggesi, An interpolating inequality for solutions of uniformly elliptic equations, In: Geometric properties for parabolic and elliptic PDE's, Cham: Springer. http://dx.doi.org/10.1007/978-3-030-73363-6_11 |
[12] | R. Magnanini, G. Poggesi, The location of hot spots and other extremal points, Math. Ann., 2021, in press. http://dx.doi.org/10.1007/s00208-021-02290-8 |
[13] |
Y. Okamoto, M. Onodera, Stability analysis of an overdetermined fourth order boundary value problem via an integral identity, J. Differ. Equations, 301 (2021), 97–111. http://dx.doi.org/10.1016/j.jde.2021.08.017 doi: 10.1016/j.jde.2021.08.017
![]() |
[14] |
G. Poggesi, Radial symmetry for p-harmonic functions in exterior and punctured domains, Appl. Anal., 98 (2019), 1785–1798. http://dx.doi.org/10.1080/00036811.2018.1460819 doi: 10.1080/00036811.2018.1460819
![]() |
[15] | G. Poggesi, The Soap Bubble Theorem and Serrin's problem: quantitative symmetry, PhD Thesis, Università di Firenze, 2019, arXiv: 1902.08584. |
[16] | J. Scheuer, Stability from rigidity via umbilicity, 2021, arXiv: 2103.07178. |
[17] |
J. Serrin, A symmetry problem in potential theory, Arch. Rational Mech. Anal., 43 (1971), 304–318. http://dx.doi.org/10.1007/BF00250468 doi: 10.1007/BF00250468
![]() |
[18] |
H. F. Weinberger, Remark on the preceding paper of Serrin, Arch. Rational Mech. Anal., 43 (1971), 319–320. http://dx.doi.org/10.1007/BF00250469 doi: 10.1007/BF00250469
![]() |
1. | Michiaki Onodera, Linear stability analysis of overdetermined problems with non-constant data, 2023, 5, 2640-3501, 1, 10.3934/mine.2023048 | |
2. | Lorenzo Cavallina, Giorgio Poggesi, Toshiaki Yachimura, Quantitative stability estimates for a two-phase Serrin-type overdetermined problem, 2022, 222, 0362546X, 112919, 10.1016/j.na.2022.112919 | |
3. | Serena Dipierro, Luca Lombardini, Partial differential equations from theory to applications: Dedicated to Alberto Farina, on the occasion of his 50th birthday, 2023, 5, 2640-3501, 1, 10.3934/mine.2023050 | |
4. | Giulio Ciraolo, Serena Dipierro, Giorgio Poggesi, Luigi Pollastro, Enrico Valdinoci, Symmetry and quantitative stability for the parallel surface fractional torsion problem, 2023, 0002-9947, 10.1090/tran/8837 | |
5. | Giorgio Poggesi, Remarks about the mean value property and some weighted Poincaré-type inequalities, 2024, 203, 0373-3114, 1443, 10.1007/s10231-023-01408-w | |
6. | Rolando Magnanini, Giorgio Poggesi, Quantitative symmetry in a mixed Serrin-type problem for a constrained torsional rigidity, 2024, 63, 0944-2669, 10.1007/s00526-023-02629-w | |
7. | Filomena Pacella, Giorgio Poggesi, Alberto Roncoroni, Optimal quantitative stability for a Serrin-type problem in convex cones, 2024, 307, 0025-5874, 10.1007/s00209-024-03555-z | |
8. | Giorgio Poggesi, Soap bubbles and convex cones: optimal quantitative rigidity, 2024, 0002-9947, 10.1090/tran/9207 | |
9. | Rolando Magnanini, Riccardo Molinarolo, Giorgio Poggesi, A General Integral Identity with Applications to a Reverse Serrin Problem, 2024, 34, 1050-6926, 10.1007/s12220-024-01693-8 |