Boundary Value Problems
VolumeΒ 2009Β (2009), Article IDΒ 185319, 19 pages
doi:10.1155/2009/185319
Research Article

A Class of 𝑝 - π‘ž -Laplacian Type Equation with Potentials Eigenvalue Problem in 𝐑 𝑁

1School of Mathematics Science, Institute of Mathematics, Nanjing Normal University, Jiangsu, Nanjing 210097, China
2College of Zhongbei, Nanjing Normal University, Jiangsu, Nanjing 210046, China

Received 20 October 2009; Accepted 6 December 2009

Academic Editor: WenmingΒ Zou

Copyright Β© 2009 Mingzhu Wu and Zuodong Yang. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Abstract

The nonlinear elliptic eigenvalue problem βˆ’ d i v ( | βˆ‡ 𝑒 | 𝑝 βˆ’ 2 βˆ‡ 𝑒 ) βˆ’ d i v ( | βˆ‡ 𝑒 | π‘ž βˆ’ 2 βˆ‡ 𝑒 ) + πœ† π‘Ž ( π‘₯ ) | 𝑒 | 𝑝 βˆ’ 2 𝑒 + πœ† 𝑏 ( π‘₯ ) | 𝑒 | π‘ž βˆ’ 2 𝑒 = 𝑓 ( π‘₯ , 𝑒 ) , 𝑒 ∈ π‘Š 1 , 𝑝 ∩ π‘Š 1 , π‘ž ( 𝐑 𝑁 ) , where 2 ≀ π‘ž ≀ 𝑝 < 𝑁 and π‘Ž ( π‘₯ ) ∈ 𝐿 𝑁 / 𝑝 ( 𝐑 𝑁 ) , 𝑏 ( π‘₯ ) ∈ 𝐿 𝑁 / π‘ž ( 𝐑 𝑁 ) , π‘Ž ( π‘₯ ) , 𝑏 ( π‘₯ ) > 0 are studied. The key ingredient is a special constrained minimization method.

1. Introduction

In this paper, we are interested in finding nontrivial weak solutions for the nonlinear eigenvalue problem ξ‚€ | | | | βˆ’ d i v βˆ‡ 𝑒 𝑝 βˆ’ 2  ξ‚€ | | | | βˆ‡ 𝑒 βˆ’ d i v βˆ‡ 𝑒 π‘ž βˆ’ 2  βˆ‡ 𝑒 + π‘Ž ( π‘₯ ) | 𝑒 | 𝑝 βˆ’ 2 𝑒 + 𝑏 ( π‘₯ ) | 𝑒 | π‘ž βˆ’ 2 𝑒 = 𝑓 ( π‘₯ , 𝑒 ) , 𝑒 ∈ π‘Š 1 , 𝑝 ∩ π‘Š 1 , π‘ž ξ€· 𝐑 𝑁 ξ€Έ , 𝑒 β‰  0 , ( 1 . 1 ) where 2 ≀ π‘ž ≀ 𝑝 < 𝑁 and π‘Ž ( π‘₯ ) ∈ 𝐿 𝑁 / 𝑝 ( 𝐑 𝑁 ) , 𝑏 ( π‘₯ ) ∈ 𝐿 𝑁 / π‘ž ( 𝐑 𝑁 ) , π‘Ž ( π‘₯ ) , 𝑏 ( π‘₯ ) > 0 , i n f π‘Ž ( π‘₯ ) , i n f 𝑏 ( π‘₯ ) β‰  0 , 𝑓 ( π‘₯ , 𝑒 ) satisfy the following conditions:

(A) 𝑓 ∈ 𝐢 ( 𝐑 𝑁 Γ— 𝐑 , 𝐑 ) , l i m 𝑑 β†’ 0 ( 𝑓 ( π‘₯ , 𝑑 ) / | 𝑑 | 𝑝 βˆ’ 1 ) = 0 , and l i m | 𝑑 | β†’ ∞ ( 𝑓 ( π‘₯ , 𝑑 ) / | 𝑑 | 𝑝 βˆ’ 1 + 𝑝 2 / 𝑁 ) = 0 uniformly in π‘₯ ∈ 𝐑 𝑁 ,(B) l i m | π‘₯ | β†’ ∞ 𝑓 ( π‘₯ , 𝑑 ) = 𝑓 ( 𝑑 ) uniformly for 𝑑 in bounded subsets of 𝑅 .

Remark 1.1. We can see if π‘Ž ( π‘₯ ) ∈ 𝐿 𝑁 / 𝑝 ( 𝐑 𝑁 ) , 𝑏 ( π‘₯ ) ∈ 𝐿 𝑁 / π‘ž ( 𝐑 𝑁 ) , then ξ€œ 𝐑 𝑁 π‘Ž ( π‘₯ ) | 𝑒 | 𝑝 ξ‚΅ ξ€œ 𝑑 π‘₯ < 𝐑 𝑁 π‘Ž ( π‘₯ ) 𝑁 / 𝑝 ξ‚Ά 1 βˆ’ 𝑝 / 𝑝 βˆ— ξ‚΅ ξ€œ 𝐑 𝑁 𝑒 𝑝 βˆ— ξ‚Ά 𝑝 / 𝑝 βˆ— ξ€œ < ∞ , 𝐑 𝑁 𝑏 ( π‘₯ ) | 𝑒 | π‘ž ξ‚΅ ξ€œ 𝑑 π‘₯ < 𝐑 𝑁 𝑏 ( π‘₯ ) 𝑁 / π‘ž ξ‚Ά 1 βˆ’ π‘ž / π‘ž βˆ— ξ‚΅ ξ€œ 𝐑 𝑁 𝑒 π‘ž βˆ— ξ‚Ά π‘ž / π‘ž βˆ— < ∞ , ( 1 . 2 ) where 𝑝 βˆ— = 𝑁 𝑝 / ( 𝑁 βˆ’ 𝑝 ) and π‘ž βˆ— = 𝑁 π‘ž / ( 𝑁 βˆ’ π‘ž ) .

Problem (1.1) comes, for example, from a general reaction-diffusion system: 𝑒 𝑑 [ ] = d i v 𝐷 ( 𝑒 ) βˆ‡ 𝑒 + 𝑐 ( π‘₯ , 𝑒 ) , ( 1 . 3 ) where 𝐷 ( 𝑒 ) = ( | βˆ‡ 𝑒 | 𝑝 βˆ’ 2 + | βˆ‡ 𝑒 | π‘ž βˆ’ 2 ) . This system has a wide range of applications in physics and related sciences such as biophysics, plasma physics, and chemical reaction design. In such applications, the function 𝑒 describes a concentration, the first term on the right-hand side of (1.3) corresponds to the diffusion with a diffusion coefficient 𝐷 ( 𝑒 ) ; whereas the second one is the reaction and relates to source and loss processes. Typically, in chemical and biological applications, the reaction term 𝑐 ( π‘₯ , 𝑒 ) is a polynomial of 𝑒 with variable coefficients.

When 𝑝 = π‘ž = 2 , problem (1.1) is a normal Schrodinger equation which has been extensively studied, for example, [18]. The authors used many different methods to study the equation. In [8], the authors established some embedding results of weighted Sobolev spaces of radially symmetric functions which are used to obtain ground state solutions. In [6], the authors studied the equation depending upon the local behavior of 𝑉 near its global minimum. In [3], the authors used spectral properties of the Schrodinger operator to study nonlinear Schrodinger equations with steep potential well. In [9], the author imposed on functions π‘˜ and 𝐾 conditions ensuring that this problem can be written in a variational form. We know that π‘Š 1 , 𝑝 ( 𝐑 𝑁 ) is not a Hilbert space for 1 < 𝑝 < 𝑁 , except for 𝑝 = 2 . The space π‘Š 1 , 𝑝 ( 𝐑 𝑁 ) with 𝑝 β‰  2 does not satisfy the Lieb lemma (e.g., see [9]). And 𝐑 𝑁 results in the loss of compactness. So there are many difficulties to study equation (1.1) of 𝑝 = π‘ž β‰  2 by the usual methods. There seems to be little work on the case 𝑝 = π‘ž β‰  2 for problem (1.1), to the best of our knowledge. In this paper, we overcome these difficulties and study (1.1) of 𝑝 β‰₯ π‘ž β‰₯ 2 .

Recently, when 𝑝 = π‘ž , π‘Ž ( π‘₯ ) = 𝑏 ( π‘₯ ) , and 𝑓 ( π‘₯ , 𝑒 ) = 0 then the problem is the following eigenvalue problem has been studied by many authors:

ξ‚€ | | | | βˆ’ d i v βˆ‡ 𝑒 𝑝 βˆ’ 2  βˆ‡ 𝑒 = 𝑉 ( π‘₯ ) | 𝑒 | 𝑝 βˆ’ 2 𝑒 , 𝑒 ∈ 𝐷 0 1 , 𝑝 ( Ξ© ) , 𝑒 β‰  0 , ( 1 . 4 ) where Ξ© βŠ† 𝐑 𝑁 . We can see [1013]. In [13], Szulkin and Willem generalized several earlier results concerning the existence of an infinite sequence of eigenvalues.

When 𝑝 = π‘ž and π‘Ž ( π‘₯ ) , 𝑏 ( π‘₯ ) is constant then the problem is the following quasilinear elliptic equation:

ξ‚€ | | | | βˆ’ d i v βˆ‡ 𝑒 𝑝 βˆ’ 2  βˆ‡ 𝑒 + πœ† | 𝑒 | 𝑝 βˆ’ 2 𝑒 = 𝑓 ( π‘₯ , 𝑒 ) , i n Ξ© , 𝑒 ∈ π‘Š 0 1 , 𝑝 ( Ξ© ) , 𝑒 β‰  0 , ( 1 . 5 ) where 1 < 𝑝 < 𝑁 , 𝑁 β‰₯ 3 , πœ† is a parameter, Ξ© is an unbounded domain in 𝐑 𝑁 . There are many results about it we can see [1418]. Because of the unboundedness of the domain, the Sobolev compact embedding does not hold. There are many methods to overcome the difficulty. In [15], the authors used the concentration-compactness principle posed by P. L. Lions and the mountain pass lemma to solve problem (1.5). In [17, 18], the authors studied the problem in symmetric Sobolev spaces which possess Sobolev compact embedding. By the result and a min-max procedure formulated by Bahri and Li [16], they considered the existence of positive solutions of ξ‚€ | | | | βˆ’ d i v βˆ‡ 𝑒 𝑝 βˆ’ 2  βˆ‡ 𝑒 + 𝑒 𝑝 βˆ’ 1 = π‘ž ( π‘₯ ) 𝑒 𝛼 i n 𝐑 𝑁 , ( 1 . 6 ) where π‘ž ( π‘₯ ) satisfies some conditions. We can see if πœ† is function, then it cannot easily be proved by the above methods.

When π‘Ž ( π‘₯ ) , 𝑏 ( π‘₯ ) is positive constant, He and Li used the mountain pass theorem and concentration-compactness principle to study the following elliptic problem in [19]: ξ‚€ | | | | βˆ’ d i v βˆ‡ 𝑒 𝑝 βˆ’ 2  ξ‚€ | | | | βˆ‡ 𝑒 βˆ’ d i v βˆ‡ 𝑒 π‘ž βˆ’ 2  βˆ‡ 𝑒 + π‘š | 𝑒 | 𝑝 βˆ’ 2 𝑒 + 𝑛 | 𝑒 | π‘ž βˆ’ 2 𝑒 = 𝑓 ( π‘₯ , 𝑒 ) i n 𝐑 𝑁 , 𝑒 ∈ π‘Š 1 , 𝑝 ∩ π‘Š 1 , π‘ž ξ€· 𝐑 𝑁 ξ€Έ , ( 1 . 7 ) where π‘š , 𝑛 > 0 , 𝑁 β‰₯ 3 , and 1 < π‘ž < 𝑝 < 𝑁 , 𝑓 ( π‘₯ , 𝑒 ) / 𝑒 𝑝 βˆ’ 1 tends to a positive constant 𝑙 as 𝑒 β†’ + ∞ . The authors prove in this paper that the problem possesses a nontrivial solution even if the nonlinearity 𝑓 ( π‘₯ , 𝑑 ) does not satisfy the Ambrosetti-Rabinowitz condition.

In [20], Li and Liang used the mountain pass theorem to study the following elliptic problem: ξ‚€ | | | | βˆ’ d i v βˆ‡ 𝑒 𝑝 βˆ’ 2  ξ‚€ | | | | βˆ‡ 𝑒 βˆ’ d i v βˆ‡ 𝑒 π‘ž βˆ’ 2  βˆ‡ 𝑒 + | 𝑒 | 𝑝 βˆ’ 2 𝑒 + | 𝑒 | π‘ž βˆ’ 2 𝑒 = 𝑓 ( π‘₯ , 𝑒 ) i n 𝐑 𝑁 , 𝑒 ∈ π‘Š 1 , 𝑝 ∩ π‘Š 1 , π‘ž ξ€· 𝐑 𝑁 ξ€Έ , ( 1 . 8 ) where 1 < π‘ž < 𝑝 < 𝑁 . They generalized a similar result for 𝑝 -Laplacian type equation in [15].

It is our purpose in this paper to study the existence of ground state to the problem (1.1) in 𝐑 𝑁 . We call any minimizer a ground state for (1.1). We inspired by [9, 16, 21] try to use constrained minimization method to study problem (1.1). Let us point out that although the idea was used before for other problems, the adaptation to the procedure to our problem is not trivial at all. But since both 𝑝 - and π‘ž -Laplacian operators are involved, careful analysis is needed. A typical difficulty for problem (1.1) in 𝐑 𝑁 is the lack of compactness of the Sobolev imbedding due to the invariance of 𝐑 𝑁 under the translations and rotations. However, our method has essential difference with the methods used in [19, 20]. In order to obtain the results, we have to overcome two main difficulties; one is that 𝐑 𝑁 results in the loss of compactness; the other is that π‘Š 1 , 𝑝 ( 𝐑 𝑁 ) is not a Hilbert space for 1 < 𝑝 < 𝑁 and it does not satisfy the Lieb lemma, except for 𝑝 = 2 .

The paper is organized as follows. In Section 2, we state some condition and many lemmas which we need in the proof of the main theorem. In Section 3, we give the proof of the main result of the paper.

2. Preliminaries

Let ξ€œ 𝐹 ( π‘₯ , 𝑑 ) = 𝑑 0 𝑓 ( π‘₯ , 𝑠 ) 𝑑 𝑠 , ξ€œ 𝐹 ( 𝑑 ) = 𝑑 0 𝑓 ( 𝑠 ) 𝑑 𝑠 ( 2 . 1 ) and we define variational functionals 𝐼 ∢ π‘Š 1 , 𝑝 ∩ π‘Š 1 , π‘ž ( 𝐑 𝑁 ) β†’ 𝑅 and 𝐼 ∞ ∢ π‘Š 1 , 𝑝 ∩ π‘Š 1 , π‘ž ( 𝐑 𝑁 ) β†’ 𝑅 by

1 𝐼 ( 𝑒 ) = 𝑝 ξ€œ 𝐑 𝑁 | | | | βˆ‡ 𝑒 𝑝 1 𝑑 π‘₯ + π‘ž ξ€œ 𝐑 𝑁 | | | | βˆ‡ 𝑒 π‘ž ξ€œ 𝑑 π‘₯ βˆ’ 𝐑 𝑁 𝐼 𝐹 ( π‘₯ , 𝑒 ) 𝑑 π‘₯ , ∞ 1 ( 𝑒 ) = 𝑝 ξ€œ 𝐑 𝑁 | | | | βˆ‡ 𝑒 𝑝 1 𝑑 π‘₯ + π‘ž ξ€œ 𝐑 𝑁 | | | | βˆ‡ 𝑒 π‘ž ξ€œ 𝑑 π‘₯ βˆ’ 𝐑 𝑁 𝐹 ( 𝑒 ) 𝑑 π‘₯ . ( 2 . 2 )

Solutions to problem (1.1) will be found as minimizers of the variational problem

𝐼 πœ† ξ‚» = i n f 𝐼 ( 𝑒 ) ; 𝑒 ∈ π‘Š 1 , 𝑝 ξ€· 𝐑 𝑁 ξ€Έ , ξ€œ 𝐑 𝑁 π‘Ž ( π‘₯ ) | 𝑒 | 𝑝 + 𝑏 ( π‘₯ ) | 𝑒 | π‘ž ξ‚Ό 𝑑 π‘₯ = πœ† , πœ† > 0 . ( 𝐼 πœ† )

To find a solution of problem ( 𝐼 πœ† ) we introduce the (limit) variational problem

𝐼 ∞ πœ† ξ‚» 𝐼 = i n f ∞ ( 𝑒 ) ; 𝑒 ∈ π‘Š 1 , 𝑝 ξ€· 𝐑 𝑁 ξ€Έ , ξ€œ 𝐑 𝑁 π‘Ž ( π‘₯ ) | 𝑒 | 𝑝 + 𝑏 ( π‘₯ ) | 𝑒 | π‘ž ξ‚Ό 𝑑 π‘₯ = πœ† , πœ† > 0 . ( 𝐼 ∞ πœ† )

Lemma 2.1. Let ( 𝑒 𝑛 ) βŠ† π‘Š 0 1 , 𝑝 ( Ξ© ) a bounded sequence and 𝑝 β‰₯ 2 . Going if necessary to a subsequence, one may assume that 𝑒 𝑛 ⇀ 𝑒 in π‘Š 0 1 , 𝑝 ( Ξ© ) , 𝑒 𝑛 β†’ 𝑒 almost everywhere, where Ξ© βŠ† 𝐑 𝑁 is an open subset.
Then,
l i m 𝑛 β†’ ∞ ξ€œ Ξ© | | βˆ‡ 𝑒 𝑛 | | 𝑝 𝑑 π‘₯ β‰₯ l i m 𝑛 β†’ ∞ ξ€œ Ξ© | | βˆ‡ 𝑒 𝑛 | | βˆ’ βˆ‡ 𝑒 𝑝 𝑑 π‘₯ + l i m 𝑛 β†’ ∞ ξ€œ Ξ© | | | | βˆ‡ 𝑒 𝑝 𝑑 π‘₯ . ( 2 . 3 )

Proof. When 𝑝 = 2 from Brezis-Lieb lemma (see [21, Lemma 1 . 3 2 ]) we have l i m 𝑛 β†’ ∞ ξ€œ Ξ© | | βˆ‡ 𝑒 𝑛 | | 2 𝑑 π‘₯ = l i m 𝑛 β†’ ∞ ξ€œ Ξ© | | βˆ‡ 𝑒 𝑛 | | βˆ’ βˆ‡ 𝑒 2 𝑑 π‘₯ + l i m 𝑛 β†’ ∞ ξ€œ Ξ© | | | | βˆ‡ 𝑒 2 𝑑 π‘₯ , ( 2 . 4 ) when 3 β‰₯ 𝑝 > 2 , using the lower semicontinuity of the 𝐿 𝑝 -norm with respect to the weak convergence and 𝑒 𝑛 ⇀ 𝑒 in π‘Š 1 , 𝑝 ( Ξ© ) , we deduce  | | βˆ‡ 𝑒 𝑛 | | 𝑝 βˆ’ 2 βˆ‡ 𝑒 𝑛 , βˆ‡ 𝑒 𝑛 ξ‚­ β‰₯  | | | | βˆ‡ 𝑒 𝑝 βˆ’ 2 ξ‚­ βˆ‡ 𝑒 , βˆ‡ 𝑒 + π‘œ ( 1 ) , l i m 𝑛 β†’ ∞  | | βˆ‡ 𝑒 𝑛 | | βˆ’ βˆ‡ 𝑒 𝑝 βˆ’ 2 βˆ‡ 𝑒 𝑛 , βˆ‡ 𝑒 𝑛 ξ‚­ β‰₯ l i m 𝑛 β†’ ∞  | | βˆ‡ 𝑒 𝑛 | | βˆ’ βˆ‡ 𝑒 𝑝 βˆ’ 2 βˆ‡ 𝑒 𝑛 ξ‚­ , βˆ‡ 𝑒 = l i m 𝑛 β†’ ∞  | | βˆ‡ 𝑒 𝑛 | | βˆ’ βˆ‡ 𝑒 𝑝 βˆ’ 2 βˆ‡ 𝑒 , βˆ‡ 𝑒 𝑛 ξ‚­ = l i m 𝑛 β†’ ∞  | | βˆ‡ 𝑒 𝑛 | | βˆ’ βˆ‡ 𝑒 𝑝 βˆ’ 2 ξ‚­ . βˆ‡ 𝑒 , βˆ‡ 𝑒 ( 2 . 5 ) Then, l i m 𝑛 β†’ ∞ ξ€œ Ξ© ξ€· | | βˆ‡ 𝑒 𝑛 | | 𝑝 βˆ’ | | | | βˆ‡ 𝑒 𝑝 ξ€Έ 𝑑 π‘₯ = l i m 𝑛 β†’ ∞ ξ€œ Ξ© | | βˆ‡ 𝑒 𝑛 | | 𝑝 βˆ’ 2 ξ‚€ | | βˆ‡ 𝑒 𝑛 | | 2 βˆ’ | | | | βˆ‡ 𝑒 2  𝑑 π‘₯ + l i m 𝑛 β†’ ∞ ξ€œ Ξ© ξ‚€ | | βˆ‡ 𝑒 𝑛 | | 𝑝 βˆ’ 2 βˆ’ | | | | βˆ‡ 𝑒 𝑝 βˆ’ 2  | | | | βˆ‡ 𝑒 2 𝑑 π‘₯ = l i m 𝑛 β†’ ∞ ξ€œ Ξ© ξ‚€ | | βˆ‡ 𝑒 𝑛 | | 𝑝 βˆ’ 2 + | | | | βˆ‡ 𝑒 𝑝 βˆ’ 2 | |  ξ‚€ βˆ‡ 𝑒 𝑛 | | 2 βˆ’ | | | | βˆ‡ 𝑒 2  𝑑 π‘₯ + l i m 𝑛 β†’ ∞ ξ€œ Ξ© ξ‚€ | | βˆ‡ 𝑒 𝑛 | | 𝑝 βˆ’ 2 | | | | βˆ‡ 𝑒 2 βˆ’ | | | | βˆ‡ 𝑒 𝑝 βˆ’ 2 | | βˆ‡ 𝑒 𝑛 | | 2  𝑑 π‘₯ . ( 2 . 6 ) From 𝑒 𝑛 ⇀ 𝑒 in π‘Š 1 , 𝑝 ( Ξ© ) , l i m 𝑛 β†’ ∞ ξ€œ Ξ© ξ‚€ | | βˆ‡ 𝑒 𝑛 | | 𝑝 βˆ’ 2 | | | | βˆ‡ 𝑒 2 βˆ’ | | | | βˆ‡ 𝑒 𝑝 βˆ’ 2 | | βˆ‡ 𝑒 𝑛 | | 2  . 𝑑 π‘₯ = 0 ( 2 . 7 ) So l i m 𝑛 β†’ ∞ ξ€œ Ξ© ξ€· | | βˆ‡ 𝑒 𝑛 | | 𝑝 βˆ’ | | | | βˆ‡ 𝑒 𝑝 ξ€Έ 𝑑 π‘₯ = l i m 𝑛 β†’ ∞ ξ€œ Ξ© ξ‚€ | | βˆ‡ 𝑒 𝑛 | | 𝑝 βˆ’ 2 + | | | | βˆ‡ 𝑒 𝑝 βˆ’ 2 | |  ξ‚€ βˆ‡ 𝑒 𝑛 | | 2 βˆ’ | | | | βˆ‡ 𝑒 2  𝑑 π‘₯ β‰₯ l i m 𝑛 β†’ ∞ ξ€œ Ξ© | | βˆ‡ 𝑒 𝑛 | | βˆ’ βˆ‡ 𝑒 𝑝 βˆ’ 2 ξ‚€ | | βˆ‡ 𝑒 𝑛 | | 2 βˆ’ | | | | βˆ‡ 𝑒 2  . ( 2 . 8 ) So we have  | | βˆ‡ 𝑒 𝑛 | | 𝑝 βˆ’ 2 βˆ‡ 𝑒 𝑛 , βˆ‡ 𝑒 𝑛 ξ‚­ +  | | βˆ‡ 𝑒 𝑛 | | βˆ’ βˆ‡ 𝑒 𝑝 βˆ’ 2 βˆ‡ 𝑒 , βˆ‡ 𝑒 𝑛 ξ‚­ +  | | βˆ‡ 𝑒 𝑛 | | βˆ’ βˆ‡ 𝑒 𝑝 βˆ’ 2 βˆ‡ 𝑒 𝑛 ξ‚­ β‰₯  | | , βˆ‡ 𝑒 βˆ‡ 𝑒 𝑛 | | βˆ’ βˆ‡ 𝑒 𝑝 βˆ’ 2 βˆ‡ 𝑒 𝑛 , βˆ‡ 𝑒 𝑛 ξ‚­ +  | | βˆ‡ 𝑒 𝑛 | | βˆ’ βˆ‡ 𝑒 𝑝 βˆ’ 2 ξ‚­ +  | | | | βˆ‡ 𝑒 , βˆ‡ 𝑒 βˆ‡ 𝑒 𝑝 βˆ’ 2 ξ‚­ βˆ‡ 𝑒 , βˆ‡ 𝑒 + π‘œ ( 1 ) . ( 2 . 9 ) Then,  | | βˆ‡ 𝑒 𝑛 | | 𝑝 βˆ’ 2 βˆ‡ 𝑒 𝑛 , βˆ‡ 𝑒 𝑛 ξ‚­ β‰₯  | | βˆ‡ 𝑒 𝑛 | | βˆ’ βˆ‡ 𝑒 𝑝 βˆ’ 2 βˆ‡ 𝑒 𝑛 βˆ’ βˆ‡ 𝑒 , βˆ‡ 𝑒 𝑛 ξ‚­ +  | | | | βˆ’ βˆ‡ 𝑒 βˆ‡ 𝑒 𝑝 βˆ’ 2 ξ‚­ βˆ‡ 𝑒 , βˆ‡ 𝑒 + π‘œ ( 1 ) l i m 𝑛 β†’ ∞ ξ€œ Ξ© | | βˆ‡ 𝑒 𝑛 | | 𝑝 𝑑 π‘₯ β‰₯ l i m 𝑛 β†’ ∞ ξ€œ Ξ© | | βˆ‡ 𝑒 𝑛 | | βˆ’ βˆ‡ 𝑒 𝑝 𝑑 π‘₯ + l i m 𝑛 β†’ ∞ ξ€œ Ξ© | | | | βˆ‡ 𝑒 𝑝 𝑑 π‘₯ , ( 2 . 1 0 ) when 𝑝 > 3 , there exists a π‘˜ ∈ 𝑁 that 0 < 𝑝 βˆ’ π‘˜ ≀ 1 . Then, we only need to prove the following inequality: l i m 𝑛 β†’ ∞ ξ€œ Ξ© ξ€· | | βˆ‡ 𝑒 𝑛 | | 𝑝 βˆ’ | | | | βˆ‡ 𝑒 𝑝 ξ€Έ 𝑑 π‘₯ β‰₯ l i m 𝑛 β†’ ∞ ξ€œ Ξ© | | βˆ‡ 𝑒 𝑛 | | βˆ’ βˆ‡ 𝑒 𝑝 βˆ’ π‘˜ ξ‚€ | | βˆ‡ 𝑒 𝑛 | | π‘˜ βˆ’ | | | | βˆ‡ 𝑒 π‘˜  . ( 2 . 1 1 )
The proof of it is similar to the above, so we omit it here. So, the lemma is proved.

Lemma 2.2. Let { 𝑒 𝑛 } be a bounded sequence in π‘Š 1 , 𝑝 ( 𝐑 𝑁 ) such that l i m 𝑛 β†’ ∞ s u p 𝑦 ∈ 𝐑 𝑁 ξ€œ 𝐡 ( 𝑦 , 𝑅 ) 𝑒 π‘ž 𝑛 𝑑 π‘₯ = 0 , 𝑝 ≀ π‘ž < 𝑝 βˆ— ( 2 . 1 2 ) for some 𝑅 > 0 . Then 𝑒 𝑛 β†’ 0 in 𝐿 𝑠 ( 𝐑 𝑁 ) for 𝑝 < 𝑠 < 𝑝 βˆ— , where 𝑝 βˆ— = 𝑁 𝑝 / ( 𝑁 βˆ’ 𝑝 ) .

Proof. We consider the case 𝑁 β‰₯ 3 . Let π‘ž < 𝑠 < 𝑝 βˆ— and 𝑒 ∈ π‘Š 1 , 𝑝 ( 𝐑 𝑁 ) . Holder and Sobolev inequalities imply that | 𝑒 | 𝐿 𝑠 ( 𝐡 ( 𝑦 , 𝑅 ) ) ≀ | 𝑒 | 𝐿 1 βˆ’ πœ† π‘ž ( 𝐡 ( 𝑦 , 𝑅 ) ) | 𝑒 | πœ† 𝐿 𝑝 βˆ— ( 𝐡 ( 𝑦 , 𝑅 ) ) ≀ 𝐢 | 𝑒 | 𝐿 1 βˆ’ πœ† π‘ž ( 𝐡 ( 𝑦 , 𝑅 ) ) ξ‚Έ ξ€œ ( 𝐡 ( 𝑦 , 𝑅 ) ) ξ€· | 𝑒 | 𝑝 + | | | | βˆ‡ 𝑒 𝑝 ξ€Έ ξ‚Ή πœ† / 𝑝 , ( 2 . 1 3 ) where πœ† = ( ( 𝑠 βˆ’ π‘ž ) / ( 𝑝 βˆ— βˆ’ π‘ž ) ) ( 𝑝 βˆ— / 𝑠 ) . Choosing πœ† = 𝑝 / 𝑠 , we obtain ξ€œ 𝐡 ( 𝑦 , 𝑅 ) | 𝑒 | 𝑠 ≀ 𝐢 𝑠 | 𝑒 | 𝐿 ( 1 βˆ’ πœ† ) 𝑠 π‘ž ( 𝐡 ( 𝑦 , 𝑅 ) ) ξ€œ 𝐡 ( 𝑦 , 𝑅 ) ξ€· | 𝑒 | 𝑝 + | | | | βˆ‡ 𝑒 𝑝 ξ€Έ . ( 2 . 1 4 )
Now, covering 𝐑 𝑁 by balls of radius π‘Ÿ , in such a way that each point of 𝐑 𝑁 is contained in at most 𝑁 + 1 balls, we find
ξ€œ 𝐑 𝑁 | 𝑒 | 𝑠 ≀ ( 𝑁 + 1 ) 𝐢 𝑠 s u p 𝑦 ∈ 𝐑 𝑁 ξ‚Έ ξ€œ 𝐡 ( 𝑦 , 𝑅 ) | 𝑒 | π‘ž ξ‚Ή ( 1 βˆ’ πœ† ) 𝑠 / π‘ž ξ€œ 𝐡 ( 𝑦 , 𝑅 ) ξ€· | 𝑒 | 𝑝 + | | | | βˆ‡ 𝑒 𝑝 ξ€Έ . ( 2 . 1 5 )
Under the assumption of the lemma, 𝑒 𝑛 β†’ 0 in 𝐿 𝑠 ( 𝐑 𝑁 ) , 𝑝 < 𝑠 < 𝑝 βˆ— . The proof is complete.

Corollary 2.3. Let { 𝑒 π‘š } be a sequence in π‘Š 1 , 𝑝 ( 𝐑 𝑁 ) satisfying ∫ 0 < 𝜌 = 𝐑 𝑁 | 𝑒 π‘š | 𝑝 𝑑 π‘₯ and such that 𝑒 π‘š ⇀ 0 in π‘Š 1 , 𝑝 ( 𝐑 𝑁 ) . Then there exist a sequence { 𝑦 π‘š } βŠ‚ 𝐑 𝑁 and a function 0 β‰  𝑒 ∈ π‘Š 1 , 𝑝 ( 𝐑 𝑁 ) such that up to a subsequence 𝑒 π‘š ( β‹… + 𝑦 π‘š ) ⇀ 𝑒 in π‘Š 1 , 𝑝 ( 𝐑 𝑁 ) .

Lemma 2.4. Let 𝑓 ∈ 𝐢 ( 𝐑 𝑁 Γ— 𝐑 ) and suppose that l i m | 𝑠 | β†’ ∞ 𝑓 ( π‘₯ , 𝑠 ) | 𝑠 | 𝑝 βˆ— βˆ’ 1 = 0 ( 2 . 1 6 ) uniformly in π‘₯ ∈ 𝐑 𝑁 and | | | | ξ€· 𝑓 ( π‘₯ , 𝑠 ) ≀ 𝐢 | 𝑠 | 𝑝 βˆ’ 1 + | 𝑠 | 𝑝 βˆ— βˆ’ 1 ξ€Έ ( 2 . 1 7 ) for all π‘₯ ∈ 𝐑 𝑁 and 𝑑 ∈ 𝑅 . If 𝑒 π‘š ⇀ 𝑒 0 in π‘Š 1 , 𝑝 ( 𝐑 𝑁 ) and 𝑒 π‘š β†’ 𝑒 0 a.e. on 𝐑 𝑁 , then l i m π‘š β†’ ∞ ξ‚Έ ξ€œ 𝐑 𝑁 𝐹 ξ€· π‘₯ , 𝑒 π‘š ξ€Έ ξ€œ 𝑑 π‘₯ βˆ’ 𝐑 𝑁 𝐹 ξ€· π‘₯ , 𝑒 0 ξ€Έ ξ€œ 𝑑 π‘₯ βˆ’ 𝐑 𝑁 𝐹 ξ€· π‘₯ , 𝑒 π‘š βˆ’ 𝑒 0 ξ€Έ ξ‚Ή 𝑑 π‘₯ = 0 , ( 2 . 1 8 ) where ∫ 𝐹 ( π‘₯ , 𝑒 ) = 𝑒 0 𝑓 ( π‘₯ , 𝑑 ) 𝑑 𝑑 .

Proof. Let 𝑅 > 0 . Applying the mean value theorem we have ξ€œ 𝐑 𝑁 𝐹 ξ€· π‘₯ , 𝑒 π‘š ξ€Έ ξ€œ 𝑑 π‘₯ = | π‘₯ | ≀ 𝑅 𝐹 ξ€· π‘₯ , 𝑒 π‘š ξ€Έ ξ€œ 𝑑 π‘₯ + | π‘₯ | β‰₯ 𝑅 𝐹 ξ€· π‘₯ , 𝑒 0 + ξ€· 𝑒 π‘š βˆ’ 𝑒 0 = ξ€œ ξ€Έ ξ€Έ 𝑑 π‘₯ | π‘₯ | ≀ 𝑅 𝐹 ξ€· π‘₯ , 𝑒 π‘š ξ€Έ ξ€œ 𝑑 π‘₯ + | π‘₯ | β‰₯ 𝑅 ξ€· 𝐹 ξ€· π‘₯ , 𝑒 π‘š βˆ’ 𝑒 0 ξ€Έ ξ€· + 𝑓 π‘₯ , πœƒ 𝑒 0 + ξ€· 𝑒 π‘š βˆ’ 𝑒 0 𝑒 ξ€Έ ξ€Έ 0 ξ€Έ 𝑑 π‘₯ , ( 2 . 1 9 ) where πœƒ depends on π‘₯ and 𝑅 and satisfies 0 < πœƒ < 1 . We now write | | | | ξ€œ 𝐑 𝑁 𝐹 ξ€· π‘₯ , 𝑒 π‘š ξ€Έ ξ€œ 𝑑 π‘₯ βˆ’ 𝐑 𝑁 𝐹 ξ€· π‘₯ , 𝑒 0 ξ€Έ ξ€œ 𝑑 π‘₯ βˆ’ 𝐑 𝑁 𝐹 ξ€· π‘₯ , 𝑒 π‘š βˆ’ 𝑒 0 ξ€Έ | | | | ≀ | | | | ξ€œ 𝑑 π‘₯ | π‘₯ | ≀ 𝑅 ξ€· 𝐹 ξ€· π‘₯ , 𝑒 π‘š ξ€Έ ξ€· βˆ’ 𝐹 π‘₯ , 𝑒 0 | | | | + | | | | ξ€œ ξ€Έ ξ€Έ 𝑑 π‘₯ | π‘₯ | β‰₯ 𝑅 𝐹 ξ€· π‘₯ , 𝑒 0 ξ€Έ | | | | + | | | | ξ€œ 𝑑 π‘₯ | π‘₯ | ≀ 𝑅 𝐹 ξ€· π‘₯ , 𝑒 π‘š βˆ’ 𝑒 0 ξ€Έ | | | | + | | | | ξ€œ 𝑑 π‘₯ | π‘₯ | β‰₯ 𝑅 𝑓 ξ€· π‘₯ , πœƒ 𝑒 0 + ξ€· 𝑒 π‘š βˆ’ 𝑒 0 ξ€Έ 𝑒 0 ξ€Έ | | | | . 𝑑 π‘₯ ( 2 . 2 0 ) For each fixed 𝑅 > 0 l i m π‘š β†’ ∞ ξ€œ | π‘₯ | ≀ 𝑅 ξ€· 𝐹 ξ€· π‘₯ , 𝑒 π‘š ξ€Έ ξ€· βˆ’ 𝐹 π‘₯ , 𝑒 0 ξ€Έ ξ€Έ 𝑑 π‘₯ = 0 , l i m π‘š β†’ ∞ ξ€œ | π‘₯ | ≀ 𝑅 𝐹 ξ€· π‘₯ , 𝑒 π‘š βˆ’ 𝑒 0 ξ€Έ 𝑑 π‘₯ = 0 . ( 2 . 2 1 ) Applying (2.20) and the Holder inequality we get that | | | | ξ€œ | π‘₯ | β‰₯ 𝑅 𝑓 ξ€· π‘₯ , πœƒ 𝑒 0 + ξ€· 𝑒 π‘š βˆ’ 𝑒 0 ξ€Έ 𝑒 0 ξ€Έ | | | | ξ€œ 𝑑 π‘₯ ≀ 𝐢 | π‘₯ | β‰₯ 𝑅 ξ‚€ | | πœƒ 𝑒 0 + ξ€· 𝑒 π‘š βˆ’ 𝑒 0 ξ€Έ | | 𝑝 βˆ’ 1 | | 𝑒 0 | | + | | πœƒ 𝑒 0 + ξ€· 𝑒 π‘š βˆ’ 𝑒 0 ξ€Έ | | 𝑝 βˆ— βˆ’ 1 | | 𝑒 0 | |  ξ‚΅ ξ€œ 𝑑 π‘₯ ≀ 𝐢 | π‘₯ | β‰₯ 𝑅 | | 𝑒 0 | | 𝑝 ξ‚Ά 1 / 𝑝 ξ‚΅ ξ€œ | π‘₯ | β‰₯ 𝑅 | | πœƒ 𝑒 0 + ξ€· 𝑒 π‘š βˆ’ 𝑒 0 ξ€Έ | | 𝑝 ξ‚Ά ( 𝑝 βˆ’ 1 ) / 𝑝 ξ‚΅ ξ€œ + 𝐢 | π‘₯ | β‰₯ 𝑅 | | 𝑒 0 | | 𝑝 βˆ— ξ‚Ά 1 / 𝑝 βˆ— ξ‚΅ ξ€œ | π‘₯ | β‰₯ 𝑅 | | πœƒ 𝑒 0 + ξ€· 𝑒 π‘š βˆ’ 𝑒 0 ξ€Έ | | 𝑝 βˆ— ξ‚Ά ( 𝑝 βˆ— βˆ’ 1 ) / 𝑝 βˆ— . ( 2 . 2 2 ) Since { 𝑒 π‘š } is bounded in π‘Š 1 , 𝑝 ( 𝐑 𝑁 ) we see that l i m 𝑅 β†’ ∞ | | | | ξ€œ | π‘₯ | β‰₯ 𝑅 𝑓 ξ€· π‘₯ , πœƒ 𝑒 0 + ξ€· 𝑒 π‘š βˆ’ 𝑒 0 ξ€Έ 𝑒 0 ξ€Έ | | | | 𝑑 π‘₯ = 0 . ( 2 . 2 3 )
The result follows from (2.21) and (2.23).

Lemma 2.5. Functions 𝐼 πœ† and 𝐼 ∞ πœ† are continuous on ( 0 , ∞ ) and minimizing sequences for problems ( 𝐼 πœ† ) and ( 𝐼 ∞ πœ† ) are bounded in π‘Š 1 , 𝑝 ( 𝐑 𝑁 ) .

Proof. From condition ( 𝐴 ) , we observe that for each πœ€ > 0 there exists 𝐢 πœ€ > 0 such that | | | | , | | | | ξ€œ 𝐹 ( 𝑒 ) 𝐹 ( π‘₯ , 𝑒 ) ≀ πœ€ 𝐑 𝑁 | 𝑒 | 𝑝 ξ€œ 𝑑 π‘₯ + πœ€ 𝐑 𝑁 | 𝑒 | 𝑝 + 𝑝 2 / 𝑁 𝑑 π‘₯ + 𝐢 πœ€ ξ€œ 𝐑 𝑁 | 𝑒 | 𝛼 𝑑 π‘₯ , ( 2 . 2 4 ) where 𝑝 < 𝛼 < 𝑝 + 𝑝 2 / 𝑁 and πœ€ > 0 .
By the Holder and Sobolev inequalities we have
ξ€œ 𝐑 𝑁 | 𝑒 | 𝑝 + 𝑝 2 / 𝑁 ξ€œ 𝑑 π‘₯ = 𝐑 𝑁 | 𝑒 | 𝑝 ( 𝑝 βˆ— βˆ’ 𝑝 βˆ’ 𝑝 2 / 𝑁 ) / ( 𝑝 βˆ— βˆ’ 𝑝 ) + 𝑝 βˆ— ( 𝑝 2 / 𝑁 ) / ( 𝑝 βˆ— βˆ’ 𝑝 ) ≀ ξ‚΅ ξ€œ 𝑑 π‘₯ 𝐑 𝑁 | 𝑒 | 𝑝 ξ‚Ά ( 𝑝 βˆ— βˆ’ 𝑝 βˆ’ 𝑝 2 / 𝑁 ) / ( 𝑝 βˆ— βˆ’ 𝑝 ) ξ‚΅ ξ€œ 𝐑 𝑁 | 𝑒 | 𝑝 βˆ— ξ‚Ά 𝑝 2 / 𝑁 / ( 𝑝 βˆ— βˆ’ 𝑝 ) ≀ 𝑆 βˆ’ 1 ξ‚΅ ξ€œ 𝐑 𝑁 | 𝑒 | 𝑝 ξ‚Ά 𝑝 / 𝑁 ξ€œ 𝐑 𝑁 | | | | βˆ‡ 𝑒 𝑝 𝑑 π‘₯ , ( 2 . 2 5 ) where | 𝑒 | 𝑝 𝑝 βˆ— ≀ 𝑆 βˆ’ 1 | βˆ‡ 𝑒 | 𝑝 𝑝 .
Similarly we have
ξ€œ 𝐑 𝑁 | 𝑒 | 𝛼 ξ€œ 𝑑 π‘₯ = 𝐑 𝑁 | 𝑒 | 𝑝 ( ( 𝑝 βˆ— βˆ’ 𝛼 ) / ( 𝑝 βˆ— βˆ’ 𝑝 ) ) + 𝑝 βˆ— ( ( 𝛼 βˆ’ 𝑝 ) / ( 𝑝 βˆ— βˆ’ 𝑝 ) ) ≀ ξ‚΅ ξ€œ 𝑑 π‘₯ 𝐑 𝑁 | 𝑒 | 𝑝 ξ‚Ά 𝑑 π‘₯ ( 𝑝 βˆ— βˆ’ 𝛼 ) / ( 𝑝 βˆ— βˆ’ 𝑝 ) ξ‚΅ ξ€œ 𝐑 𝑁 | 𝑒 | 𝑝 βˆ— ξ‚Ά 𝑑 π‘₯ ( 𝛼 βˆ’ 𝑝 ) / ( 𝑝 βˆ— βˆ’ 𝑝 ) ≀ 𝑆 βˆ’ 𝑝 βˆ— ( 𝛼 βˆ’ 𝑝 ) / 𝑝 ( 𝑝 βˆ— βˆ’ 𝑝 ) ξ‚΅ ξ€œ 𝐑 𝑁 | 𝑒 | 𝑝 ξ‚Ά 𝑑 π‘₯ ( 𝑝 βˆ— βˆ’ 𝛼 ) / ( 𝑝 βˆ— βˆ’ 𝑝 ) ξ‚΅ ξ€œ 𝐑 𝑁 | | | | βˆ‡ 𝑒 𝑝 ξ‚Ά 𝑑 π‘₯ 𝑝 βˆ— ξ€· 𝑝 ( 𝛼 βˆ’ 𝑝 ) / 𝑝 βˆ— ξ€Έ βˆ’ 𝑝 . ( 2 . 2 6 ) Consequently by the Young inequality we have ξ€œ 𝐑 𝑁 | 𝑒 | 𝛼 ξ€œ 𝑑 π‘₯ ≀ πœ‚ 𝐑 𝑁 | | | | βˆ‡ 𝑒 𝑝 ξ‚΅ ξ€œ 𝑑 π‘₯ + 𝐾 ( πœ‚ ) 𝐑 𝑁 | 𝑒 | 𝛼 ξ‚Ά 𝑑 π‘₯ 𝑝 ( 𝑝 βˆ— βˆ’ 𝛼 ) / ( 𝑝 2 𝑝 βˆ— βˆ’ 𝑝 2 βˆ’ 𝑝 βˆ— 𝛼 ) ( 2 . 2 7 ) for πœ‚ > 0 , where 𝐾 ( πœ‚ ) > 0 is a constant.
Because 𝑒 ∈ π‘Š 1 , 𝑝 ∩ π‘Š 1 , π‘ž ( 𝐑 𝑁 ) so we can by Sobolev embedding and ∫ πœ† = 𝐑 𝑁 π‘Ž ( π‘₯ ) | 𝑒 | 𝑝 + 𝑏 ( π‘₯ ) | 𝑒 | π‘ž 𝑑 π‘₯ letting Μ‚ ∫ πœ† = 𝐑 𝑁 | 𝑒 | 𝑝 𝑑 π‘₯ < ∞ , we derive the following estimates for 𝐼 ( 𝑒 ) and 𝐼 ∞ ( 𝑒 ) :
𝐼 ( 𝑒 ) , 𝐼 ∞ ξ‚΅ 1 ( 𝑒 ) β‰₯ 𝑝 βˆ’ πœ€ 𝑆 βˆ’ 1 Μ‚ πœ† 𝑝 / 𝑁 βˆ’ 𝐢 πœ€ πœ‚ ξ‚Ά ξ€œ 𝐑 𝑁 | | | | βˆ‡ 𝑒 𝑝 + 1 𝑑 π‘₯ π‘ž ξ€œ 𝐑 𝑁 | | | | βˆ‡ 𝑒 π‘ž Μ‚ 𝑑 π‘₯ βˆ’ πœ€ πœ† βˆ’ 𝐾 ( πœ‚ ) 𝐢 πœ€ Μ‚ πœ† 𝑝 ( 𝑝 βˆ— βˆ’ 𝛼 ) / ( 𝑝 2 𝑝 βˆ— βˆ’ 𝑝 2 βˆ’ 𝑝 βˆ— 𝛼 ) . ( 2 . 2 8 )
Choosing πœ€ > 0 and πœ‚ > 0 so that
1 𝑝 βˆ’ πœ€ 𝑆 βˆ’ 1 Μ‚ πœ† 𝑝 / 𝑁 βˆ’ 𝐢 πœ€ πœ‚ > 0 , ( 2 . 2 9 ) we see that 𝐼 πœ† and 𝐼 ∞ πœ† are finite and moreover minimizing sequences for problems ( 𝐼 πœ† ) and ( 𝐼 ∞ πœ† ) are bounded. It is easy to check that 𝐼 πœ† and 𝐼 ∞ πœ† are continuous on ( 0 , ∞ ) .

We observe that 𝐼 ∞ πœ‡ ≀ 0 for all πœ‡ > 0 . Indeed, let 𝑒 ∈ 𝐢 ∞ 0 ( 𝑅 𝑁 ) and

ξ€œ 𝐑 𝑁 | | | π‘Ž ( π‘₯ ) 𝑒 ( π‘₯ / 𝜎 ) 𝜎 𝑁 / π‘ž | | | 𝑝 ξ€œ 𝑑 π‘₯ + 𝐑 𝑁 | | | 𝑏 ( π‘₯ ) 𝑒 ( π‘₯ / 𝜎 ) 𝜎 𝑁 / π‘ž | | | π‘ž 𝑑 π‘₯ = πœ‡ , ( 2 . 3 0 ) then for each 𝜎 > 0 we have

𝐼 ∞ πœ‡ ≀ 1 𝑝 𝜎 𝑝 + ( 𝑝 / π‘ž βˆ’ 1 ) 𝑁 ξ€œ 𝐑 𝑁 | | | | βˆ‡ 𝑒 𝑝 1 𝑑 π‘₯ + π‘ž 𝜎 π‘ž ξ€œ 𝐑 𝑁 | | | | βˆ‡ 𝑒 π‘ž 𝑑 π‘₯ βˆ’ 𝜎 𝑁 ξ€œ 𝐑 𝑁 𝐹 ξ€· 𝜎 βˆ’ 𝑁 / π‘ž 𝑒 ξ€Έ 𝑑 π‘₯ ⟢ 0 ( 2 . 3