On the ground state eigenfunction of a convex domain in Euclidean space Pawel Kr¨oger∗ Mathematisches Institut, Universit¨ at Erlangen-N¨ urnberg Bismarckstr. 1 1/2, 91054 Erlangen Germany

Abstract. We study the first eigenfunction φ1 of the Dirichlet Laplacian on a convex domain in Euclidean space. Elementary properties of Bessel functions yield that ||φ1 ||∞ /||φ1 ||2 → ∞ if D is a sector in Euclidean plane with area 1 and the angle tends to 0. We aim to characterize those domains D such that (vol(D))1/2 ||φ1 ||∞ /||φ1 ||2 is large by the ratio of the eigenvalues ˜ of D with given volume. of D and the infimum of the eigenvalues of all subdomains D Key: 35 J 05, 35 B 05, 35 J 25 0. Introduction. Let D be a bounded convex domain in Rd . We consider the eigenvalue problem 4φ

=

−λφ

on D

and φ|∂D ≡

0.

Let φ1 be the eigenfunction for the first eigenvalue λ1 of the above eigenvalue problem. We investigate whether ||φ1 ||∞ is bounded above by a constant multiple of ||φ1 ||2 . We use (n) (n) standard results on Bessel functions in order to show that limn→∞ ||φ1 ||∞ /||φ1 ||2 = ∞ if (n) φ1 denotes the ground state eigenfunction on a sector S (n) in Euclidean plane with angle π/n. An elementary calculation shows that there exists a sequence of rectangular domains R(n) with R(n) ⊂ S (n) such that the first eigenvalue of R(n) is ”close” to the first eigenvalue of S (n) although the ratio of the areas of R(n) and S (n) tends to 0. We will show in the second part of the paper that a related phenomenon can be used to characterize those domains in Euclidean space for which ||φ1 ||∞ /||φ1 ||2 is large. More precisely, we will show that ||φ1 ||∞ /||φ1 ||2 can be bounded above by a function of λ1 and the infimum of the first eigenvalues of all convex subdomains of D with given volume. 1. Sectors in Euclidean plane We introduce polar coordinates (r, ω) in Euclidean plane. Let Jn be the Bessel function of (n) order n and let jn be the first positive zero of Jn . The ground state eigenfunction φ1 of the Laplace operator −4 on the sector S (n) ≡ {0 < r < jn , 0 < ω < π/n} is given by (n) φ1 (r, ω) ≡ Jn (r) sin(nω) and the corresponding eigenvalue is equal to 1. (n)

Claim. Let φ1 be the ground state eigenfunction of S (n) ⊂ R2 . Then (n)

(n)

lim (area(S (n) ))1/2 ||φ1 ||∞ /||φ1 ||2 = ∞.

n→∞ ∗

Research supported by the Deutsche Forschungsgemeinschaft

1

Proof. Kapteyn’s inequality (cf. [2], p. 268) states that √ z n exp(n 1 − z 2 ) √ Jn (nz) ≤ for 0 < z < 1. (1 + 1 − z 2 )n √ We set u ≡ 1 − z 2 . Using the Taylor series of the exponential function, we can easily show exp 2u < 1 − 23 u3 for every u with 0 < u < 1. Hence, that 1−u 1+u Jn (nz) < (

(1 − u2 ) exp 2u n/2 2 ) < (1 − u3 )n/2 . 2 (1 + u) 3

2

Now suppose that z < 1 − n− 3 for a positive number . Then u = n(−1)/3 . We can conclude that  /3

n 2 2 1− Jn (nz) < (1 − n−1 )n/2 = ((1 − n−1 )3n /2 ) 3 3



1 − z2 >

 /3

< 2−n



1−z >

.

On the other hand, it follows from [2], p.488 eq. (II) and the property that the absolute values of the consecutive extremas of Jn decrease that the maximum value of Jn is greater or equal than C1 n−1/2 for an appropriate constant C1 . The value of the q independent variable 0 0 such that Jn attains its maximum will be denoted by jn . Then jn > n(n + 2) and Jn00 (r) < 0 on (n, jn0 ) (see [2], p.486 (3) and (2)). We can conclude that lim n→∞

Z

jn 1

n−n+ 3

Jn (r)2 r dr/

Z 0

jn

Jn (r)2 r dr = 1 for every  > 0.

Finally, we take into account that the smallest positive zero jn of Jn is asymptotically equal to n(1 + C2 n−2/3 ) for an appropriate constant C2 > 0 (cf. [2], p.516). This establishes the claim. Remark. A simple calculation shows that there exists a constant C1 such that S (n) contains for every n a rectangle R(n) with sides n1/3 and 1 − C1 n−2/3 . Obviously, the ratio area(R(n) )/area(S (n) ) tends to zero as n → ∞. The first Dirichlet eigenvalue of R(n) is less or equal than 1 + (1 + 3C1 )n−2/3 for n sufficiently large. On the other hand, the difference between the first zero jn of Jn and the second zero of Jn is greater or equal than C2 n1/3 for a constant C2 (this can be shown by similar methods as in [2], pp. 518–521). We can conclude that the second eigenvalue of S (n) is greater or equal than 1 + C3 n−2/3 for an appropriate constant C3 (recall that the first eigenvalue is equal to 1). This example motivates the following considerations. 2. The ground state eigenfunction and the first eigenvalues of convex domains and their subdomains ˜ ⊂ D with λD˜ ≤ λD +α(λD −λD ) It is easy to see that the existence of a (small) subdomain D 1 1 2 1 for a constant α with α < 1 yields a lower bound for ||φ1 ||∞ /||φ1 ||2 . This follows if we ˜ ˜ extend the first eigenfunction φD 1 of D by zero to D and develop the resulting function with

2

˜ We obtain that under the respect to an orthonormal system of Dirichlet eigenfunctions of D. normalization assumption ||φ1 ||2 = 1 the following holds (we will drop the superscript D): ˜

˜

˜

˜

˜

D 2 D 2 D 2 D 2 λD 1 ||φ1 ||2 ≥ λ1 < φ1 , φ1 > + λ2 (||φ1 ||2 − < φ1 , φ1 > ).

On the other hand, Z

||φ1 ||∞ ≥ (

˜ D

˜ ˜ D ˜ ≥ < φD ˜ φ21 dx)1/2 /vol(D) 1 , φ1 > /(vol(D)||φ1 ||2 )

This leads to ||φ1 ||∞ ≥



˜ 1 − α/vol(D).

We recall that Yu and Zhong have obtained the following explicit lower bound for λ2 − λ1 which might be useful if we want to apply the above observations in order to obtain a lower bound for ||φ1 ||∞ /||φ1 ||2 (see [3]). Let D be a convex domain in Rd . Let diam(D) denote the diameter of D. Then the first two Dirichlet eigenvalues λ1 and λ2 of D satisfy the inequality λ2 − λ1 ≥ π 2 (diam(D))−2 . The following theorem shows that we can also proceed in the converse direction. ˜

Theorem. Let D be a convex domain in Rd . Suppose that λD 1 ≥ Λ(δ) for every convex ˜ ˜ D ⊂ D with vol(D) ≤ δvol(D) and positive numbers Λ(δ) and δ. We normalize a positive eigenfunction φ1 for the first eigenvalue λ1 of D by ||φ1 ||2 = 1. Then ||φ1 ||∞ ≤ Cd δ −1/2 ( ln ||φ1 ||∞ − ln(1 − λ1 /Λ(δ)))d/2 for a positive constant Cd which depends only on the dimension d. ˜

Remark. We intend to choose δ not too small, say δ ≡ 1/2. We can conclude that inf{(λD 1 − 1 ˜ ˜ λ1 )/λ1 | D ⊂ D with vol(D) = 2 vol(D)} decreases exponentially fast as ||φ1 ||∞ increases. We guess that the above Theorem is no longer true if we drop the assumption that D is convex although it may be difficult to find a good lower bound for Λ(δ). Proof. Assume that vol(D) = 1. Let D() denote the set {x ∈ D| φ1 (x) > } for a positive number . We choose  such that vol(D() ) = δ. By [1], Theorem 6.1, ln φ1 is concave. This implies in particular that D() is convex. Let M denote the maximum value of φ1 . The concavity of ln φ1 yields that vol{x ∈ D | φ1 (x) ≥ λ M 1−λ } ≥ λd δ

for every 0 < λ < 1.

Hence, 1=

Z D

2

φ1 (x) dx ≥ δM

2

Z 0

1

(/M )2λ dλd ≥ δCd M 2 (ln(M/))−d

(∗)

where Cd depends only on the dimension d (the last inequality can be easily shown if we consider λ ∼ (ln(M/))−1 ). 3

On the other hand, we develop the restriction of φ1 to D() with respect an orthonormal () system of Dirichlet eigenfunctions φk of D() . The corresponding eigenalues will be denoted () by λk . By Green’s formula (n(x) stands for the exterior normal to the boundary ∂D() at x), () (λk

− λ1 )

Z D()

() φ1 φk

dx =

Z

= − =

()

D()

()

φk 4φ1 − φ1 4φk dx

Z

()

 < ∇φk , n(x) >

() ∂D Z () λk D()

()

φk dx.

If we take into account that Z ∞ X

(

k=1

and

()

D()

φ1 φk dx)2 = ||φ1 |D() ||22 ≥ δ

Z ∞ X

(

k=1

we can conclude that

()

D()

1φk dx)2 = δ

()

λ1 − λ 1 ()

λ1

≤ .

In conjunction with the assumption we arrive at 1 − λ1 /Λ(δ) ≤ . Thus, the assertion follows from (∗). References 1. H. J. BRASCAMP AND E. H. LIEB, On extensions of the Brunn-Minkowski and Pr´ekopaLeindler theorems, including inequalities for log concave functions, and with an application to the diffusion equation. J. Funct. Anal. 22 (1976), 366–389. 2. G. N. WATSON, A treatise on the theory of Bessel functions, Cambridge University Press 1922. 3. QIHUANG YU AND J. Q. ZHONG, Lower bounds of the gap between the first and second eigenvalues of the Schr¨odinger operator. Trans. Amer. Math. Soc. 294 (1986), 341–349.

4

On the ground state eigenfunction of a convex domain ...

Let φ1 be the eigenfunction for the first eigenvalue λ1 of the above eigenvalue problem. We investigate whether ||φ1||∞ is bounded above by a constant multiple ...

96KB Sizes 0 Downloads 190 Views

Recommend Documents

On the tangent cone of the convex hull of convex sets
May 14, 2010 - 1.1 Extension to finitely many convex sets. We now extend Theorem 1 to finitely many convex sets, i.e. we show the following. Corollary 4 Let k be an integer and C1,C2....,Ck be convex sets with nonempty intersection. Let z. ∗ ∈ k.

Magnetic ground state of CeNi1−xCux: A calorimetric ...
Apr 6, 2005 - (iv) The very-low-temperature CP/T behavior can show departures .... TABLE I. Crystallographic data of the CeNi1axCux compounds obtained ... tions by using grinding paper disks of different grain sizes ..... of the total R ln 2 entropy

Magnetic ground state of CeNi1−xCux: A calorimetric ...
Apr 6, 2005 - system since SR bridges the gap between neutron scatter- ing on the one side and bulk magnetic measurements on the other.45. Considering our experimental results as a whole, we can propose a convenient description of our system. Althoug

A Course on Convex Geometry
bdA boundary of A. If f is a function on Rn with values in R or in the extended real line [−∞,∞] and ..... i = 1,...,n + 2, we consider the system of linear equations.

On the contact domain method: A comparison of ...
This work focuses on the assessment of the relative performance of the so-called contact domain method, using either the Lagrange multiplier or the penalty ...

Decoherence induced deformation of the ground state in adiabatic ...
Mar 29, 2013 - 2D-Wave Systems Inc., 100-4401 Still Creek Drive, Burnaby, B.C., Canada V5C 6G9. 3Department of Physics, Simon Fraser University, Burnaby, British Columbia, Canada V5A 1S6. Despite more than a decade of research on adiabatic quantum co

A Note on Separation of Convex Sets
A line L separates a set A from a collection S of plane sets if A is contained in one of ... For any non-negative real number r, we denote by B, the disk with radius r.

Decoherence induced deformation of the ground state in adiabatic ...
Mar 29, 2013 - Decoherence induced deformation of the ground state in adiabatic quantum computation. Qiang Deng,1 Dmitri V. Averin∗,1 Mohammad H. Amin,2, 3 and Peter Smith2, 3. 1Department of Physics and Astronomy, Stony Brook University, SUNY, Sto

Excited and ground-state complexes of ...
(DNS) as electron acceptors were accompanied by the appearance of a new emission ... (CT) band was also observed between TPP and both PIC and DNS.

On the ground down under
Jun 15, 2015 - Our visit to Frasers Australand's projects in Sydney and Melbourne and ... offer a network of locations to support tenant growth and leasing or.

A Study on Convex Optimization Approaches to Image ...
ful applications include image denoising [22,8,30], image decomposition [2,17] ..... + ui/c)}) . (30). Interestingly, our experiments show that just one single step of ... experiments are computed by a Ubuntu desktop with AMD Athalon 64 X2 5600.

A Note on Convex Relaxations for Non-Metric ...
13 Aug 2012 - i,j ı{rij. 2 + sij ≤ θij} − ∑ i,j ı{pi − pj − rij = 0} −. ∑ i,j ı{sij − qi = 0}... . (3). We will derive the primal of this expression using the following variant of Fenchel duality, min x f(Ax) = max y:AT y=0. −fâ

Chiral ground-state currents of interacting ... - Research at Google
Oct 31, 2016 - come together they can construct a basic building block for creating ... A schematic illustration of how qubits and their couplers can be tiled to create a ..... 50,. 1395–1398 (1983). 15. Georgescu, I. M., Ashhab, S. & Nori, ...

A Domain Decomposition Method based on the ...
Nov 1, 2007 - In this article a new approach is proposed for constructing a domain decomposition method based on the iterative operator splitting method.

pdf-1862\introductory-lectures-on-convex-optimization-a-basic ...
Sign in. Loading… Whoops! There was a problem loading more pages. Retrying... Whoops! There was a problem previewing this document. Retrying.

A note on convex decompositions of a set of points in ...
Interchange y and x and a and b in Case 3.2.1. Subcase .... O. Aichholzer, H. Krasser, The point set order type data base: A collection of applications and results ...

Excited and ground-state complexes of ...
(DNS) as electron acceptors were accompanied by the appearance of a new emission band in .... It should be noted that the observed data show solvation.

GRADIENT ESTIMATES FOR THE GROUND STATE ...
SCHR¨ODINGER EIGENFUNCTION AND APPLICATIONS ..... [14] Yau, S.-T. Open problems in geometry, Proceedings of symposia in pure mathematics, vol.

On upper bounds for high order Neumann eigenvalues of convex ...
Let Ω be a bounded convex domain in Euclidean space Rn. By considering a manifold ... where f1/(n−1) is a positive and concave function on the interval (0,1).

The Unique Minimal Dual Representation of a Convex ...
[〈x, x∗〉 − f(x)]. When f is a convex function, there is an important duality between f and f∗ ... Example 3.1 shows that if the local Lipschitz continuity assumption.

Cutting a Convex Polyhedron Out of a Sphere
mitrescu proposed an O(log n)-approximation algorithm with O(mn + n log n) running time .... We call P to be cornered if it does not contain the center o of Q ... [Sketch only] Since P contains the center o of Q, any cutting sequence, starting from .