数学杂志  2016, Vol. 36 Issue (1): 157-163   PDF    
扩展功能
加入收藏夹
复制引文信息
加入引用管理器
Email Alert
RSS
本文作者相关文章
刘琼
含临界指数p-Kirchhoff型方程的非平凡解
刘琼     
湖北工程学院数学与统计学院, 湖北 孝感 432000
摘要:本文研究了一类含临界指数的p-Kirchhoff型方程.利用变分方法与集中紧性原理, 通过证明对应的能量泛函满足局部的(PS)c条件, 得到了这类方程非平凡解的存在性, 推广了关于Kirchhoff型方程的相关结果.
关键词p-Kirchhoff型方程    (PS)c条件    临界指数    非平凡解    
NONTRIVIAL SOLUTIONS FOR P-KIRCHHOFF TYPE EQUATIONS INVOLVING CRITICAL EXPONENT
LIU Qiong     
School of Mathematics and Statistics, Hubei Engineering University, Xiaogan 432000, China
Abstract: In this paper, we are concerned with the existence of nontrivial solutions to a class of p-Kirchhoff type equations involving critical exponent. By analyzing the effect of critical nonlinear term and estimating energy functional carefully, with the aid of variational methods and the concentration-compactness principle, we establish the existence of nontrivial solutions for the p-Kirchhoff type equations, which extends the results of the Kirchhoff type equations.
Key words: p-Kirchhoff type equation     (PS)c condition     critical exponent     nontrivial solution    
1 引言及主要结果

本文研究如下一类含Sobolev临界指数的$p$-Kirchhoff型方程

$ \left\{ \begin{array}{l} - M(\int_\Omega | \nabla u{|^p}dx)elt{a_p}u = |u{|^{{p^*} - 2}}u + \lambda f(u),\;x \in \Omega ,\\ u(x) = 0,\;x \in \partial \Omega , \end{array} \right. $ (1.1)

其中$\Omega\subset \mathbb{R}^{N}$为一光滑有界区域, ${\rm{1 < }}{p}{\rm{ < }}{N}$, ${\Delta _p}u$$p$-Laplacian算子, $p^{*}=\frac{Np}{N-p}$是Sobolev临界指数, $\lambda>0$为实参数. $M:\mathbb{R}^+\to\mathbb{R}^+$, $f:\mathbb{R}\to\mathbb{R}$为连续函数且满足如下条件:

($\mathcal{H}_{1}$)存在$M_0>0$, 使得对任意$t\geq0$有, $M(t)\geq M_0$;

($\mathcal{H}_{2}$)存在$\theta>\frac{p}{p^*}$, 使得对任意$t\geq0$有, $\widetilde{M}(t)\geq\theta M(t)t$, 其中$\widetilde{M}(t)=\int_0^tM(s)ds$;

($\mathcal{H}_{3}$) $\mathop {\lim }\limits_{|u| \to 0} \frac{{f(u)}}{{|u{|^{p - 1}}}} = 0$;

($\mathcal{H}_{4}$)存在$q\in(p, p^{*})$, 使得$\mathop {\lim }\limits_{|u| \to + \infty } \frac{{f(u)}}{{|u{|^{q - 1}}}} = 0$.

($\mathcal{H}_{5}$)存在$\mu\in(\frac{p}{\theta}, p^{*})$, 使得对任意$u\neq0$有, $0<\mu F(u)\leq uf(u)$, 其中$F(u)=\int_0^uf(s)ds$, $\theta$为($\mathcal{H}_{2}$)中所给出.

近年来, Kirchhoff型方程

$ \left\{ \begin{array}{l} -(a + b\int_\Omega | \nabla u{|^2}dx)eltau = f(x, u), \;x \in \Omega, \\ u = 0, \;x \in \partial \Omega, \end{array} \right. $ (1.2)

由于其广泛的应用而引起了许多数学研究者的注意.例如文[1, 2]研究了方程(1.2) 正解的存在性; 文[3, 4]得到了方程(1.2) 无穷多个正解的存在性; 文[5]研究了方程(1.2) 非平凡解与变号解的存在性; 文[6]考虑了方程(1.2) 多重非平凡解的存在性.其它相关的结果见文献[7, 8, 9, 10].本文考虑更一般的含Sobolev临界指数的$p$-Kirchhoff型方程(1.1), 讨论方程(1.1) 非平凡解的存在性.由于方程(1.1) 包含临界指数使得嵌入$W_0^{1, p}(\Omega)\hookrightarrow L^{p^{*}}(\Omega)$非紧, 这给我们利用变分方法带来了困难.我们将利用P.L.Lions的集中紧性原理(文[11])通过对$\left( {{\rm{PS}}} \right)$序列的仔细分析, 选择恰当的能量水平使得局部的${\left( {{\rm{PS}}} \right)_c}$条件成立.

定义最佳常数

$ \mathcal{S} = \mathop {\inf }\limits_{u \in W_0^{1, p}(\Omega )\backslash \{ 0\} } \frac{{\int_\Omega | \nabla u{|^p}dx}}{{{{(\int_\Omega | u{|^{{p^*}}}dx)}^{\frac{p}{{{p^*}}}}}}}. $ (1.3)

本文的主要结果为:

定理 1.1  若条件($\mathcal{H}_{1}$)—($\mathcal{H}_{5}$)满足.则存在$\Lambda>0$, 使得当$\lambda\geq\Lambda$时, 方程(1.1) 至少存在一个能量水平于$\Big(0, (\frac1\mu-\frac1{p^*})(M_0 \mathcal{S})^{\frac{N}{p}}\Big)$的非平凡解.

注 1  若令$M(t)=a+bt$, 其中$a, b$为正的常数.则$M(t)\geq a$

$ \widetilde{M}(t)=\int_{0}^{t}M(s)ds=\int_{0}^{t}(a+bs) ds=at+\frac{1}{2}bt^{2}\geq \frac{1}{2}(a+bt)t=\theta M(t)t, $

其中$\theta=\frac{1}{2}$.从而条件$(\mathcal{H}_{1}$)--($\mathcal{H}_{2}$)满足.若令$f(u)=\sum\limits_{i=1}^{k}|u|^{q_{i}-2}u$, 其中$k \ge 1,2p < {q_i} < {p^*}$, 易知条件$(\mathcal{H}_{3}$)--($\mathcal{H}_{5}$)满足.

2 主要结果的证明

在整篇文章中, $C, C_{i}$表示正的常数它们在不同的行或段中可以不同; “$\rightarrow$”表示强收敛, “$\rightharpoonup$”表示弱收敛; $W_{0}^{1, p}(\Omega)$表示通常的Sobolev空间, 其范数为$\|u\|=(\int_\Omega|\nabla u|^pdx)^{\frac{1}{p}}$; $|u|_{p}:=\big(\int_{\Omega}|u|^{p}dx\big)^{\frac{1}{p}}$表示Lebesgue空间$L^{p}(\Omega)$的范数.

方程(1.1) 对应的能量泛函定义为

$ \begin{equation}\label{E21} \mathcal{E}(u)=\frac1p\widetilde{M}(\|u\|^p)-\frac1{p^*}\int_\Omega |u|^{p^*}dx-\lambda\int_\Omega F(u)dx, \end{equation} $ (2.1)

由条件($\mathcal{H}_{1}$)-($\mathcal{H}_{5}$)易知$\mathcal{E}(u)\in C^{1}(W_{0}^{1, p}(\Omega), \mathbb{R})$, 从而方程(1.1) 的非平凡解等价于能量泛函$\mathcal{E}(u)$$W_0^{1, p}(\Omega)$上的非零临界点.

首先我们证明能量泛函$\mathcal{E}(u)$具有山路几何.

引理2.1  若($\mathcal{H}_{1}$)--($\mathcal{H}_{5}$)成立, 则

(ⅰ)存在$\rho, \alpha>0$使得$\mathcal{E}(u)\big|_{\|u\|=\rho}\geq\alpha>0$.

(ⅱ)对任意$\lambda>0$, 存在独立于$\lambda$的非负函数$u_{0}\in W_{0}^{1, p}(\Omega)$满足$\|u_{0}\|>\rho$, 使得$\mathcal{E}(u_{0})<0$.

  (ⅰ)由($\mathcal{H}_{3}$)与($\mathcal{H}_{4}$)知, 对任意$\varepsilon>0$, 存在$C_{\varepsilon}>0$, 使得对任意$u\in \mathbb{R}$

$ \begin{equation}\label{E22} F(u)\leq\frac\varepsilon p |u|^p+\frac{C_{\varepsilon}}{q}|u|^q. \end{equation} $ (2.2)

根据($\mathcal{H}_{1}$)及Sobolev嵌入定理可得

$ \begin{array}{l} \mathcal{E}(u) \ge \frac{{{M_0}}}{p}{\left\| u \right\|^p} - {C_1}{\left\| u \right\|^{{p^*}}} - \lambda {C_2}\varepsilon {\left\| u \right\|^p} - \lambda {C_3}{\left\| u \right\|^q}\\ \;\;\;\;\;\;\;\;\; = (\frac{{{M_0}}}{p} - \lambda {C_2}\varepsilon ){\left\| u \right\|^p} - {C_1}{\left\| u \right\|^{{p^*}}} - \lambda {C_3}{\left\| u \right\|^q}. \end{array} $

$\varepsilon=\frac{M_0}{2p\lambda C_2}$, 则

$ \begin{equation} \mathcal{E}(u)\geq\frac{M_0}{2p}\|u\|^{p}-C_1\|u\|^{p^{*}}-\lambda C_3\|u\|^{q}. \end{equation} $ (2.3)

由于$q\in(p, p^*)$, 从而由(2.3) 式知, 存在$\rho>0, \alpha>0$使得$\mathcal{E}(u)\big|_{\|u\|=\rho}\geq\alpha$.

(ⅱ)取非负函数$\phi_0\in C_0^\infty(\Omega)$且满足$\|\phi_0\|=1$.结合条件($\mathcal{H}_{2}$), 对任意$t\geq t_0>0$可得

$ \begin{equation}\label{E24} \widetilde{M}(t)\leq \frac{\widetilde{M}(t_0)}{t_0^{\frac{1}{\theta}}}t^{\frac{1}{\theta}} =C_0t^{\frac{1}{\theta}}. \end{equation} $ (2.4)

由($\mathcal{H}_{5}$)知, $\int_\Omega F(t\phi_0)dx\geq0$.因此对任意$t\geq t_0$

$ \mathcal{E}(t\phi_0)\leq\frac{C_0}pt^{\frac{p}{\theta}} -\frac{t^{p^{*}}}{p^{*}}\int_\Omega \phi_0^{p^{*}}dx. $

又由于$\frac{p}{\theta } < {p^*}$, 所以可以选取$t^*>0$充分大, 并令$u_{0}=t^*\phi_0$则可满足引理要求.

引理2.2  若$\{u_n\}\subset W_0^{1, p}(\Omega)$$\mathcal{E}$${\left( {{\rm{PS}}} \right)_c}$序列, 则$\{u_n\}$$W_0^{1, p}(\Omega)$中有界.

  设$\{u_n\}\subset W_0^{1, p}(\Omega)$$\mathcal{E}$${\left( {{\rm{PS}}} \right)_c}$序列, 即

$ \begin{equation}\label{E28} \mathcal{E}(u_n)\to c, \quad \quad \mathcal{E}'(u_n)\to 0, \end{equation} $ (2.5)

从而由($\mathcal{H}_{5}$)对充分大的$n$, 及($\mathcal{H}_{1}$)与($\mathcal{H}_{2}$)可得

$ \begin{array}{l} c + {o_n}(1) \ge \mathcal{E}({u_n}) - \frac{1}{\mu }\langle \mathcal{E}'({u_n}),{u_n}\rangle \\ \;\;\;\;\;\;\;\;\;\;\;\;\;\; \ge \frac{1}{p}\widetilde M({\left\| u \right\|^p}) - \frac{1}{\mu }M({\left\| u \right\|^p}){\left\| u \right\|^p} \ge (\frac{\theta }{p} - \frac{1}{\mu }){M_0}{\left\| u \right\|^p}, \end{array} $ (2.6)

又由于$\mu>\frac{p}{\theta}$, 所以$\{u_n\}$$W_0^{1, p}(\Omega)$中有界.

引理2.3  若$\{u_n\}\subset W_0^{1, p}(\Omega)$$\mathcal{E}$${\left( {{\rm{PS}}} \right)_c}$序列, 若$c<(\frac1\mu-\frac1{p^*})(M_0 \mathcal{S})^{\frac{N}{p}}$, 则$\{u_n\}$$W_0^{1, p}(\Omega)$中存在强收敛的子列.

  由引理2.2知$\{u_n\}$有界, 因此, 可以假设(必要时可以取子列)

$ \left\{ \begin{array}{l} {u_n} \rightharpoonup u\;\;\;在W_0^{1,p}(\Omega ),\\ {u_n} \to u\;\;\;{\rm{ }}{在}{L^s}(\Omega ),1 \le s < {p^*},\\ {u_n} \to u\;\;\;{\rm{a}}{\rm{.e}}{\rm{.}}{在}\Omega ,\\ |\nabla {u_n}{|^p} \rightharpoonup \vartheta ,\\ |{u_n}{|^{{p^*}}} \rightharpoonup \nu , \end{array} \right. $ (2.7)

其中$\vartheta$$\nu$$\overline{\Omega}$上非负有界测度.则由P.L.Lions集中紧性原理(文[11])知, 存在至多可数集$J$使得

$ \left\{ {\begin{array}{*{20}{l}} {\nu = |u{|^{{p^*}}} + \sum\limits_{j \in J} {{\nu _j}} {\delta _x}_j,\;\;\;\;{\nu _j} > 0,}\\ {\vartheta \ge |\nabla u{|^p} + \sum\limits_{j \in J} {{\vartheta _j}} {\delta _{{x_j}}},\;\;\;\;{\vartheta _j} > 0,}\\ {\mathcal{S}\nu _j^{\frac{p}{{{p^*}}}} \le {\vartheta _j},} \end{array}} \right. $ (2.8)

这里${\delta _{{x_j}}}$为Dirac测度集中在点$x_j\in\overline{\Omega}$.

$\psi(x)\in C_0^\infty(\Omega)$使得$0\leq\psi(x)\leq1$; $\psi(x)=1$, 若$|x|<1$; $\psi(x)=0$, 若$|x|\geq2$, 且$|\nabla\psi|\leq 2$.对$\varepsilon>0$$j\in J$, 记$\psi_\varepsilon^j(x)=\psi((x-x_j)/\varepsilon)$.由于$\mathcal{E}'(u_n)\to 0$$\{\psi_\varepsilon^ju_n\}$有界, $\langle \mathcal{E}'(u_n), \psi_\varepsilon^ju_n\rangle\to0$ ($n\to\infty$), 所以有

$ \begin{array}{l} \;\;\;M({\left\| {{u_n}} \right\|^p})\int_\Omega | \nabla {u_n}{|^p}\psi _\varepsilon ^jdx\\ = - M({\left\| {{u_n}} \right\|^p})\int_\Omega {{u_n}} |\nabla {u_n}{|^{p - 2}}\nabla {u_n}\nabla \psi _\varepsilon ^jdx\\ \;\;\; + \int_\Omega | {u_n}{|^{{p^*}}}\psi _\varepsilon ^jdx + \lambda \int_\Omega f ({u_n}){u_n}\psi _\varepsilon ^jdx + {o_n}(1). \end{array} $ (2.9)

由(2.8)式与Vitali定理, 可得

$ \mathop {\lim }\limits_{n \to \infty } \int_\Omega | {u_n}\nabla \psi _\varepsilon ^j{|^p}dx = \int_\Omega | u\nabla \psi _\varepsilon ^j{|^p}dx, $

因此, 由Hölder不等式可得

$ \begin{array}{*{20}{l}} {\mathop {\lim \sup }\limits_{n \to \infty } |\int_\Omega {{u_n}} |\nabla {u_n}{|^{p - 2}}\nabla {u_n}\nabla \psi _\varepsilon ^jdx|}\\ { \le \mathop {\lim \sup }\limits_{n \to \infty } {{(\int_\Omega | \nabla {u_n}{|^p}dx)}^{\frac{{p - 1}}{p}}}{{(\int_\Omega | {u_n}\nabla \psi _\varepsilon ^j{|^p}dx)}^{\frac{1}{p}}}}\\ { \le {C_4}{{(\int_{B({x_j},2\varepsilon )} | u{|^p}|\nabla \psi _\varepsilon ^j{|^p}dx)}^{\frac{1}{p}}}}\\ { \le {C_4}{{(\int_{B({x_j},2\varepsilon )} | \nabla \psi _\varepsilon ^j{|^N}dx)}^{\frac{1}{N}}}{{(\int_{B({x_j},2\varepsilon )} | u{|^{{p^*}}}dx)}^{\frac{1}{{{p^*}}}}}}\\ { \le {C_5}{{(\int_{B({x_j},2\varepsilon )} | u{|^{{p^*}}}dx)}^{\frac{1}{{{p^*}}}}} \to 0\;\;\;(\varepsilon \to 0)\;.} \end{array} $ (2.10)

另一方面, 由(2.8)式可得$f({u_n}){u_n} \to f(u){u_n}\;\;\;{\rm{a}}.{\rm{e}}.$$\Omega$中,且$u_n\to u$$L^p(\Omega)$$L^q(\Omega)$中.由($\mathcal{H}_{3}$)--($\mathcal{H}_{5}$), 对任意$\varepsilon>0$, 存在$C_\varepsilon>0$, 使得

$ \begin{equation}\label{E215} |f(t)|\leq\varepsilon|t|^{p-1}+C_\varepsilon|t|^{q-1}, \quad \forall \ t\in\mathbb{R}, \end{equation} $ (2.11)

因此$|f(u_n)u_n|\leq\varepsilon|u_n|^p+C_\varepsilon|u_n|^q.$由此应用Vitali定理, 则有

$ \mathop {\lim }\limits_{n \to \infty } \int_\Omega f ({u_n}){u_n}dx = \int_\Omega f (u)udx. $

由于$\psi_\varepsilon^j$具有紧支集, 在(2.9)式中令$n\to\infty$, 由(2.8) 与(2.10)式可得

$ M_0\int_\Omega\psi_\varepsilon^jd\mu\leq C_5\Big(\int_{B(x_j, 2\varepsilon)}|u|^{p^{*}} dx\Big)^{\frac{1}{p^{*}}} +\lambda\int_{B(x_j, 2\varepsilon)} f(u)udx+\int_\Omega\psi_\varepsilon^jd\nu. $

$\varepsilon\to0$, 可得$M_0\vartheta_j\leq\nu_j$.因此

$ \begin{equation}\label{E216} (M_0 \mathcal{S})^{\frac{N}{p}}\leq\nu_j. \end{equation} $ (2.12)

下证以上不等式不可能成立.假设对某$j_0\in J$$(M_0 \mathcal{S})^{\frac{N}{p}}\leq\nu_{j_0}$.由($\mathcal{H}_{2}$)可得

$ \frac1p\widetilde{M}(\|u_n\|^p)-\frac1\mu M(\|u_n\|^p)\|u_n\|^p\geq0. $

$\{u_{n}\}$$\mathcal{E}$${\left( {{\rm{PS}}} \right)_c}$序列知

$ c=\mathcal{E}(u_n)-\frac1\mu \langle \mathcal{E}'(u_n), u_n\rangle+o_n(1), $

从而有$c \ge (\frac{1}{\mu } - \frac{1}{{{p^*}}})\int_\Omega | {u_n}{|^{{p^*}}}dx + {o_n}(1) \ge (\frac{1}{\mu } - \frac{1}{{{p^*}}})\int_\Omega {\psi _\varepsilon ^{{j_0}}} |{u_n}{|^{{p^*}}}dx + {o_n}(1),{\rm{ }}$$n\to\infty$, 可得

$ c\geq \big(\frac1\mu-\frac1{p^{*}}\big)\sum\limits_{j\in J}\psi_\varepsilon^{j_0}(x_j)\nu_j \geq \big(\frac1\mu-\frac1{p^{*}}\big)(M_0 \mathcal{S})^{\frac{N}{p}}. $

这与$c<\big(\frac1\mu-\frac1{p^{*}}\big)(M_0 \mathcal{S})^{\frac{N}{p}}$矛盾, 所以$J=\emptyset$, 从而$u_n\to u$$L^{p^{*}}(\Omega)$中.由(2.11)式可得

$ \begin{array}{l} \int_\Omega | f({u_n})({u_n} - u)|dx \le \int_\Omega ( \varepsilon |{u_n}{|^{p - 1}} + {C_\varepsilon }|{u_n}{|^{q - 1}})|{u_n} - u|dx\\ \;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\; \le \varepsilon {(\int_\Omega | {u_n}{|^p}dx)^{\frac{{p - 1}}{p}}}{(\int_\Omega | {u_n} - u{|^p}dx)^{\frac{1}{p}}}\\ \;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\; + {C_\varepsilon }{(\int_\Omega | {u_n}{|^q}dx)^{\frac{{q - 1}}{q}}}{(\int_\Omega | {u_n} - u{|^q}dx)^{\frac{1}{q}}}. \end{array} $

因此再应用(2.8) 式, 可得

$ \mathop {\lim }\limits_{n \to \infty } \int_\Omega f ({u_n})({u_n} - u)dx = 0. $ (2.13)

由于$u_n\to u$$L^{p^{*}}(\Omega)$中, 可得

$ \mathop {\lim }\limits_{n \to \infty } \int_\Omega | {u_n}{|^{{p^*} - 2}}{u_n}({u_n} - u)dx = 0. $ (2.14)

$\langle \mathcal{E}'(u_n), u_n-u\rangle=o_n(1)$, 可推得

$ \begin{array}{l} \langle \mathcal{E}'({u_n}),{u_n} - u\rangle = M({\left\| {{u_n}} \right\|^p})\int_\Omega | \nabla {u_n}{|^{p - 2}}\nabla {u_n}\nabla ({u_n} - u)dx\\ \;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\; - \int_\Omega | {u_n}{|^{{p^*} - 2}}{u_n}({u_n} - u)dx - \lambda \int_\Omega f ({u_n})({u_n} - u)dx = {o_n}(1), \end{array} $

由上式与(2.13), (2.14) 式可得

$ \mathop {\lim }\limits_{n \to \infty } M({u_n}{^p})\int_\Omega | \nabla {u_n}{|^{p - 2}}\nabla {u_n}\nabla ({u_n} - u)dx = 0. $

由于$\{u_n\}$是有界的且$M$连续, 必要时取子列, 则存在$t_0\geq0$使得

$ M(\|u_n\|^p)\to M(t_0^p)\geq M_0, \quad n\to\infty, $

从而$\underset{n\to\infty}\lim\int_\Omega|\nabla u_n|^{p-2}\nabla u_n\nabla(u_n-u)dx=0.$因此由$(S_+)$性质(见文献[12]定义2.3) 有, $u_n\to u$$W_0^{1, p}(\Omega)$中.引理2.3证毕.

$ {c^*} = \mathop {\inf }\limits_{\gamma \in \Gamma } {\max _{t \in [0,1]}}\mathcal{E}(\gamma (t)) > 0, $ (2.15)

其中$\Gamma=\big\{\gamma\in C([0,1], W_0^{1, p}(\Omega)): \gamma(0)=0, \, \mathcal{E}(\gamma(1))<0\big\}.$由引理2.1, 根据无$\left( {{\rm{PS}}} \right)$条件的山路引理知, 存在序列$\{u_n\}\subset W_0^{1, p}(\Omega)$使得$\mathcal{E}(u_n)\to c^{*},\quad\text{且} \ \mathcal{E}'(u_n)\to 0.$

引理2.4  若($\mathcal{H}_{1}$)—($\mathcal{H}_{5}$)成立.则存在$\Lambda>0$, 使得对任意$\lambda\geq\Lambda$

$ c^{*}\in \big(0, (\frac1\mu-\frac1{p^*})(M_0 \mathcal{S})^{\frac{N}{p}}\big), $

其中$c^{*}$由(2.15) 式给出.

  设$u_{0}$由引理2.1(ⅱ)给出, 易知$\lim\limits_{t\to+\infty} \mathcal{E}(tu_{0})=-\infty$, 从而存在$t_\lambda>0$使得

$ \mathcal{E}(t_\lambda u_{0})=\underset{t\geq0}\max \mathcal{E}(tu_{0}). $

因此

$ t_\lambda^{p-1}M(\|t_\lambda u_{0}\|^p)\|u_{0}\|^p =\lambda \int_\Omega f(t_\lambda u_{0})u_{0}\, dx +t_\lambda^{p^{*}-1}\int_\Omega |u_{0}|^{p^{*}}dx, $

$ \begin{equation}\label{E27} M(\|t_\lambda u_{0}\|^p)\|t_\lambda u_{0}\|^p =\lambda t_\lambda \int_\Omega f(t_\lambda u_{0})u_{0}\, dx +t_\lambda^{p^{*}}\int_\Omega |u_{0}|^{p^{*}}dx. \end{equation} $ (2.16)

由(2.3) 可得, 当${t_0} < {t_\lambda }$时有

$ \frac{C_0}\theta\|u_{0}\|^{\frac{p}{\theta}}t_\lambda^{\frac{p}{\theta}} \geq t_\lambda^{p^{*}}\int_\Omega |u_{0}|^{p^{*}}dx. $

又因为$\frac{p}{\theta } < {p^*}$, 从而$t_\lambda$有界.因此存在序列$\lambda_n\to+\infty$使得$t_{\lambda_n}\to T\geq0$ ($n\to\infty$).因此, 存在$C>0$, 使得对任意的$n$$M(\|t_{\lambda_n}u_{0}\|^p)\|t_{\lambda_n}u_{0}\|^p\leq C$, 即

$ \lambda_n t_{\lambda_n}\int_\Omega f(t_{\lambda_n} u_{0})u_{0}\, dx +t_{\lambda_n}^{p^{*}}\int_\Omega |u_{0}|^{p^{*}}dx\leq C. $

$T>0$, 由上面的不等式可知, 当$n\to\infty$时有

$ + \infty \leftarrow {\lambda _n}{t_{{\lambda _n}}}\int_\Omega f ({t_{{\lambda _n}}}{u_0}){u_0}{\mkern 1mu} dx + t_{{\lambda _n}}^{{p^*}}\int_\Omega | {u_0}{|^{{p^*}}}dx \le C, $

矛盾, 因此$T=0$.令$\gamma^*(t)=tu_{0}$.显然$\gamma^*\in\Gamma$, 则有

$ 0 < {c^*} \le \mathop {\max }\limits_{t \ge 0} {\rm{ }}\mathcal{E}({\gamma ^*}(t)) = \mathcal{E}({t_\lambda }{u_0}) \le \frac{1}{p}\widetilde M({\left\| {{t_\lambda }{u_0}} \right\|^{p}}). $

$t_{\lambda_n}\to0$$(\frac1\mu-\frac1{p^{*}})(M_0 \mathcal{S})^{\frac{N}{p}}>0$知, 当$\lambda>0$充分大时, $ \frac1p\widetilde{M}(\|t_\lambda u_{0}\|^p)<\big(\frac1\mu-\frac1{p^{*}}\big)(M_0 \mathcal{S})^{\frac{N}{p}}, $进而有$ 0 < {c^*} < (\frac{1}{\mu } - \frac{1}{{{p^*}}}){({M_0}\mathcal{S})^{\frac{N}{p}}}. $引理2.4证毕.

定理1.1的证明  由引理2.1, 引理2.2, 引理2.3, 引理2.4及山路引理(见文献[13, 14])知,$\mathcal{E}$必有临界点$u\in W_{0}^{1, p}(\Omega)$, 满足$\mathcal{E}(u)=c^{*}\in\Big(0, \big(\frac1\mu-\frac1{p^{*}}\big)(M_0 \mathcal{S})^{\frac{N}{p}}\Big)$,即$u$是方程(1.1) 的一个非平凡解且能量水平位于$\Big(0, \big(\frac1\mu-\frac1{p^{*}}\big)(M_0 \mathcal{S})^{\frac{N}{p}}\Big)$.定理1.1证毕.

参考文献
[1] Alves C O, Corrêa F J S A, Ma T F. Positive solutions for a quasilinear elliptic equation of Kirchhoff type[J]. Comput. Math. Appl., 2005, 49: 85–93. DOI:10.1016/j.camwa.2005.01.008
[2] Cheng B T, Wu X. Existence results of positive solutions of Kirchhoff problems[J]. Nonl. Anal., 2009, 71: 4883–4892. DOI:10.1016/j.na.2009.03.065
[3] He X M, Zou W M. Infinitely many positive solutions for Kirchhoff-type problems[J]. Nonl. Anal., 2009, 70: 1407–1414. DOI:10.1016/j.na.2008.02.021
[4] Liu D C. On a p-Kirhhoff equation via fountain theorem and dual fountain theorem[J]. Nonl. Anal., 2010, 72: 208–302.
[5] Perera K, Zhang Z T. Nontrival solutions of Kirchhoff-type problems via the Yang index[J]. J. Diff.Equ, 2006, 221: 246–255. DOI:10.1016/j.jde.2005.03.006
[6] Sun J, Tang C L. Existence and multiplicity of solutions for Kirchhoff type equations[J]. Nonl. Anal., 2011, 74: 1212–1222. DOI:10.1016/j.na.2010.09.061
[7] Figueiredo G M. Existence of a positive solution for a Kirchhoff problem type with critical growth via truncation argument[J]. J. Math. Anal. Appl., 2013, 401: 706–713. DOI:10.1016/j.jmaa.2012.12.053
[8] Cheng B T. New existence and multiplicity of nontrivial solutions for nonlocal elliptic Kirchhoff type problems[J]. J. Math. Anal. Appl., 2012, 394: 488–495. DOI:10.1016/j.jmaa.2012.04.025
[9] Lü D F. A note on Kirchhoff-type equations with Hartree-type nonlinearities[J]. Nonl. Anal., 2014, 99: 35–48. DOI:10.1016/j.na.2013.12.022
[10] Ma T F. Remarks on an elliptic equation of Kirchhoff type[J]. Nonl. Anal., 2005, 63: 1967–1977. DOI:10.1016/j.na.2005.03.021
[11] Lions P L. The concentration-compactness principle in the calculus of variations, the limit case[J]. Rev. Mat. Iberoam, 1985, 1: 145–201.
[12] De Nápoli P, Mariani M C. Mountain pass solutions to equations of p-Laplacian type[J]. Nonlinear Anal., 2003, 54: 1205–1219. DOI:10.1016/S0362-546X(03)00105-6
[13] Rabinowitz P. Minimax methods in critical points theory with applications to differential equations[M]. CBMS Reg. Conf. Ser. Math., Vol. 65, Providence, RI: Amer. Math. Soc., 1986.
[14] Willem M. Minimax theorems[M]. Boston: Birkhäuser, 1996.