Bifurcation of symmetric solutions for the sublinear Moore-Nehari differential equation (Q831506)

In this paper, the author considers the bifurcation of symmetric nodal solutions for the sublinear Moore-Nehari differential equation \[u^{\prime\prime}+h\left( x,\lambda\right) \left\vert u\right\vert ^{p-1}u=0,~u\left( -1\right) =u\left( 1\right) =0,\tag{1.1} \] where \(0<p<1,\) \(h\left( x,\lambda\right) =0\) for \(\left\vert x\right\vert <\lambda\) and \(h\left( x,\lambda\right) =1\) for \(\lambda\leq\left\vert x\right\vert \leq1\) and \(\lambda\in\left( 0,1\right) \) is a bifurcation parameter. Denote the set of all solutions \(\left( \lambda,u\right) \) for (1.1) by \(S\) and define \(S_{n}:=\left\{ \left( \lambda,u_{n}\left( x,\lambda\right) \right) :~0<\lambda<1\right\} .\) Then the author gives the following main theorem: Theorem. Let \(0<p<1\) and let \(n\) be a non-negative integer. If \(n\) is even, then \(S_{n}\) does not bifurcate. Let \(n\) be an odd integer. Then there exists a unique number \(\lambda_{\ast}\left( n\right) \in\left( 0,1\right) \) such that \(S_{n}\) bifurcates at \(\lambda=\lambda_{\ast}\left( n\right) \) only and it does not have any other bifurcation point. We denote the set of all solutions for (1.1) by \(S\) and put \(\lambda_{\ast}:=\lambda_{\ast}\left( n\right) \) and \(u_{\ast}:=u_{n}\left( x,\lambda_{\ast}\left( n\right) \right) \). Then there exists a subset \(C_{n}\) of \(S\) satisfying the conditions below: (i) \(C_{n}\) bifurcates from \(S_{n}\) at \((\lambda_{\ast},u_{\ast})\). (ii) \(C_{n}\) is closed, connected and bounded in \((0,1)\times C_{0} ^{1}[-1,1].\) (iii) \(C_{n}\) \(\cup\) \(S_{n}\) is a connected component of \(S\) having \((\lambda_{\ast},u_{\ast})\). (iv) \(C_{n}\) \(\cap\) \(S_{n}=\{(\lambda_{\ast},u_{\ast})\}.\) (v) Put \(n=2m-1\) with \(m\geq1\). Let \(C_{m,m-1}\left( C_{m-1,m}\right) \) denote the set of \((\lambda,u)\in C_{2m-1}\) such that \(u\) is an \((m,m-1)\)-solution (\((m-1,m)\)-solution, respectively). Then it holds that \(\ C_{2m-1}=C_{m,m-1}\cup C_{m-1,m}\cup\{(\lambda_{\ast},u_{\ast})\}\). Therefore, each point in \(C_{2m-1}\backslash\) \(\{(\lambda_{\ast},u_{\ast})\}\) is either an \((m,m-1)\)-solution or an \((m-1,m)\)-solution of (1.1). (vi) If \((\lambda,u)\in C_{2m-1}\), then \((\lambda,-u(-x))\in C_{2m-1}\). Hence, \((\lambda,u)\) is in \(C_{m,m-1}\) if and only if \((\lambda,-u(-x))\) lies in \(C_{m-1,m}\). (vii) For any \(\lambda\in(\lambda_{\ast},1)\), there exist points \(u,v\) such that \((\lambda,u)\in C_{m,m-1}\) and \((\lambda,v)\in\) \(C_{m-1,m}\). (viii) For any \((\lambda,u)\in\) \(C_{n}\), \(u^{\prime}\left( 0\right) \) is positive. (ix) \(\left\Vert u\right\Vert _{C^{1}}\rightarrow0\) as \(\lambda\rightarrow1\) with \((\lambda,u)\in S\). In particular, this assertion is valid for \((\lambda,u)\in\) \(C_{n}\) also. The author gives some preliminary lemmas and properties for the proof and then proves the above theorem.