Genericity of nondegenerate geodesics with general boundary conditions (Q1958996)

Let \(M\) be a possibly noncompact manifold of dimension \(n\). For each \(\nu\in\{0,\dots,n\}\), consider the set \(\text{Met}_{\nu}^{k}(M)\) of all semi-Riemannian \(C^{k}\) metrics of index \(\nu\), which is a subset of \(\Gamma_{\text{sym}}^{k}(T^{*}M\otimes T^{*}M)\), the set of \(C^{k}\) sections \(s\) such that \(s_{x}:T_{x}M\times T_{x}M\to\mathbb R\) is symmetric for all \(x\in M\). If \(M\) is compact, \(\Gamma_{\text{sym}}^{k}(T^{*}M\otimes T^{*}M)\) has a natural Banach space structure, and \(\text{Met}_{\nu}^{k}(M)\) is an open subset. In order to endow the space of tensors over a noncompact manifold \(M\) with a Banach space structure the authors, consider a vector subspace \(\mathcal E\) of \(\Gamma_{\text{sym}}^{k}(T^{*}M\otimes T^{*}M)\), called a \(C^{k}\)-Whitney type Banach space of tensor fields over \(M\), satisfying the conditions: {\parindent=6mm \begin{itemize}\item[(a)] \(\mathcal E\) contains all tensor fields having compact support; \item[(b)] \(\mathcal E\) is endowed with a Banach space norm \(\|\cdot\|_{\mathcal E}\) with the property that \(\|\cdot\|_{\mathcal E}\)-convergence of a sequence implies convergence in the weak Whitney \(C^{k}\)-topology. \end{itemize}} Using an auxiliary Riemannian metric \(g_{\text{R}}\) on \(M\), it is possible to construct a \(C^{k}\)-Whitney type Banach space of tensors on \(M\). Such a space \(\mathcal E\) will be fixed in the sequel. Let \(\mathcal A_{\nu}\subset\mathcal E\cap\text{Met}_{\nu}^{k}(M)\) be an open subset of the intersection. Let \(\mathcal M\) be the product \(M\times M\). For all semi-Riemannian metrics \(g\) on \(M\) define \(\bar g=g\oplus(-g)\) on \(T\mathcal M=TM\oplus TM\). For all submanifolds \(\mathcal P\subset \mathcal M\), define \(\mathcal A_{\nu,\mathcal P}=\{g\in\mathcal A_{\nu}\,:\, \mathcal P \text{ is nondegenerate with respect to }\bar g=g\oplus(-g)\}\). If \(\mathcal P\) is compact, then \(\mathcal A_{\nu,\mathcal P}\) is open in \(\mathcal A_{\nu}\). A \(\nu\)-general boundary condition on \(M\) is a compact submanifold \(\mathcal P\subset M\times M\) such that \(\mathcal A_{\nu,\mathcal P}\) is nonempty. It is well-known that the set \(H^{1}([0,1],M)\) of all curves of Sobolev class \(H^{1}\) in \(M\) has a canonical Hilbert manifold structure. The subset \[ \Omega_{\mathcal P}(M)=\{\gamma\in H^{1}([0,1],M)\,:\, (\gamma(0),\gamma(1))\in\mathcal P\} \] is a submanifold of \(H^{1}([0,1],M)\). Fix \(g\in\mathcal A_{\nu,\mathcal P}\). A \(g\)-geodesic \(\gamma\in\Omega_{\mathcal P}(M)\) is a \((g,\mathcal P)\)-geodesic if it satisfies \((\dot\gamma(0),\dot\gamma(1))\in (T_{(\gamma(0),\gamma(1))}\mathcal P)^{\perp}\), where \(\perp\) denotes orthogonality relative to \(\bar g\). A \((g,\mathcal P)\)-geodesic \(\gamma\) is a critical point of the generalized energy \(C^{k}\)-functional \(f:\mathcal A_{\nu,\mathcal P}\times\Omega_{\mathcal P}(M)\to\mathbb R\) defined by \(f(g,\gamma)=\frac{1}{2}\int_{0}^{1}g(\dot\gamma,\dot\gamma)\,dt\). A \((g,\mathcal P)\)-geodesic is nondegenerate if it is a nondegenerate critical point of \(f\). A \(\nu\)-general boundary condition \(\mathcal P\) is admissible if, for every \(g_{0}\in\mathcal A_{\nu,\mathcal P}\), there exists an open neighbourhood \(\mathcal V\) of \(g_{0}\) in \(\mathcal A_{\nu,\mathcal P} \) and \(a>0\), such that for all \(g\in\mathcal V\), an every \((g,\mathcal P)\)-geodesic \(\gamma\), the \(g_{\text{R}}\)-length \(L_{\text{R}}(\gamma)\geq a\). With this definitions the main result of the paper is the following: Theorem 5.10: Let \(M\) be a smooth \(n\)-dimensional manifold and \(\nu\in\{0,\dots,n\}\) an index. Fix \(\mathcal E\subset\Gamma_{\text{sym}}^{k}(T^{*}M\otimes T^{* }M)\) a \(C^{k}\)-Whitney type Banach space of tensor fields over \(M\) and \(\mathcal A_{\nu}\subset\mathcal E\cap\text{Met}_{\nu}^{k}(M)\) an open subset. Consider \(\mathcal P\) an admissible \(\nu\)-general boundary condition. Then the following is a generic subset in \(\mathcal A_{\nu,\mathcal P}\) \[ \mathcal G_{\mathcal P}(M)=\{g\in \mathcal A_{\nu,\mathcal P}\,:\, \text{ all } (g,\mathcal P)\text{-geodesics }\gamma\in\Omega_{\mathcal P}(M)\text{ are nondegenerate}\}. \] The paper ends with an extension of the above genericity in the \(C^{\infty}\)-context. Renaming \(\mathcal A_{\nu,\mathcal P}\) and \(\mathcal G_{\mathcal P}(M)\) as \(\mathcal A_{\nu,\mathcal P}^{k}\) and \(\mathcal G_{\mathcal P}^{k}(M)\) to stress the dependence on \(C^{k}\)-regularity of tensor fields, the authors show that \(\mathcal G_{\mathcal P}^{\infty}(M):=\bigcap_{k\in\mathbb N}\mathcal G_{ \mathcal P}^{k}(M)\) is generic in \(\mathcal A_{\nu,\mathcal P}^{\infty}=\bigcap_{k\in\mathbb N}\mathcal A_{\nu,\mathcal P}^{k}\) in the \(C^{\infty}\)-topology.