The Krein formula for generalized resolvents in degenerated inner product spaces (Q1291096)

Let \({\mathcal H}\) be a, possibly degenerated, inner product space such that \({\mathcal H}/({\mathcal H}\cap{\mathcal H}^{\perp})\) is a Pontryagin space, and let \(S\) be a symmetric operator in \({\mathcal H}\) with defect index \((1,1)\). If \({\mathcal P}\) is a Pontryagin space containing \({\mathcal H}\) as a singular subspace, and \(A\) is a selfadjoint relation in \({\mathcal P}\) extending \(S\), we call the set of analytic functions \[ [(A- z)^{-1}x,y],\quad x,y\in{\mathcal H}\tag{1} \] a generalized resolvent of \(S\). In the following, denote by \(\langle\cdot\rangle\) the linear span of the elements between the brackets. If \(A\) is \({\mathcal H}\)-minimal in the sense that \[ \overline{\langle{\mathcal H},(A- z)^{-1}{\mathcal H}|z\in\varrho(A)\rangle}={\mathcal P}, \] the corresponding generalized resolvent is called minimal. In this case, if \({\mathcal P}\) has index \(\kappa\) of negativity \((\kappa= \text{ind}_-{\mathcal P})\), the generalized resolvent is said to have index \(\kappa\). If \(\tau\) is a complex valued meromorphic function, denote by \(\varrho(\tau)\) its domain of holomorphy. For \(\kappa\in \mathbb{N}_0\) let the set \({\mathcal N}_{\kappa}\) of generalized Nevanlinna functions \(\tau\) be defined as follows: \(\tau\) is meromorphic in \(\mathbb{C}\setminus\mathbb{R}\), \(\tau(\overline z)= \overline{\tau(z)}\) for \(z\in\varrho(\tau)\), and the Nevanlinna kernel \((z,w\in\varrho(\tau))\), \[ N_\tau(z, w)= \begin{cases} {\tau(z)- \tau(\overline w)\over z-\overline w},\quad & z\neq\overline w\\ \tau'(z),\quad & z=\overline w\end{cases} \] has \(\kappa\) negative squares. If \({\mathcal H}\) is nondegenerated, i.e., a Pontryagin space, Kreĭn's formula \[ [(A- z)^{-1}x,y]= [(\mathring A- z)^{-1}x,y]- [x,\chi(\overline z)]{1\over \tau(z)+ q(z)} [\chi(z), y]\tag{2} \] establishes a bijective correspondence between the minimal generalized resolvents of index \(\kappa\) of \(S\) and the set \({\mathcal N}_{\kappa- \kappa_0}(\kappa_0= \text{ind}_-{\mathcal H})\) of parameters \(\tau\). Here \(\mathring A\subseteq{\mathcal H}^2\) is a fixed selfadjoint extension of \(S\), \(\chi(z)\) are certain defect elements of \(S\) and \(q(z)\) is a corresponding \(Q\)-function, which means that the relation \(N_q(z,w)= [\chi(z),\chi(w)]\) holds. In this paper it is considered that \({\mathcal H}\) is actually degenerated, \(\dim{\mathcal H}^0= \Delta>0\). It turns out that the relation (2) still holds, where \(\mathring A\), \(\chi(z)\) and \(q(z)\) have a similar meaning. However, the parameter \(\tau\) does not run through the Nevanlinna class \({\mathcal N}_{\kappa- \kappa_0}\), but through a different set of functions \({\mathcal K}^\Delta_{\kappa- \kappa_0}\) which is defined as follows: Definition. For \(\nu, \Delta\in \mathbb{N}_0\), denote by \({\mathcal K}^\Delta_\nu\) the set of all complex valued functions \(\tau(z)\), meromorphic in \(\mathbb{C}\setminus\mathbb{R}\), which satisfy \(\tau(\overline z)= \overline{\tau(z)}\) for \(z\in\varrho(\tau)\), and are such that the maximal number of the negative squares of the quadratic forms \((m\in \mathbb{N}_0,z_1,\dots, z_m\in \varrho(\tau))\) \[ Q(\xi_1,\dots, \xi_m; \eta_0,\dots, \eta_{\Delta- 1})= \sum^n_{i,j= 1} N_\tau(z_i, z_j)\xi_i\overline{\xi_j}+ \sum^{\Delta- 1}_{k= 0} \sum^m_{i= 1} \text{Re}(z^k_i\xi_i \overline{\eta_k})\tag{3} \] is \(\nu\).