Conditions for sectoriality and compactness of the resolvent for a non-self-adjoint Sturm–Liouville operator with singular distributional potential
Sergey N. Tumanov · 2025
Read Conditions for sectoriality and compactness of the resolvent for a non-self-adjoint Sturm–Liouville operator with singular distributional potential by Sergey N. Tumanov (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2503-16172.txt.
Introduction
Tumanov presents a 2025 paper on a Sturm-Liouville operator whose potential is rougher than a plain function, and which lacks self-adjointness. Two traits are put on trial: whether the operator is sectorial, and whether its resolvent is compact. For each trait the paper gives conditions that are both needed and enough. The tests are phrased through the potential's own antiderivative, a function one step smoother than the potential itself.
Catalog record
| CREATOR | Sergey N. Tumanov |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 68 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2503-16172
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Conditions for sectoriality and compactness of the resolvent for a non-self-adjoint Sturm–Liouville operator with singular distributional potential Sergey N. Tumanov Abstract The aim of this paper is to find necessary and sufficient conditions for sectoriality and compactness of the resolvent for Sturm–Liouville operators with complex-valued potentials of the class q ∈ W 2 , l o c − 1 ( ℝ + ) q\in W_{2,loc}^{-1}(\mathbb{R}_{+}) in terms of its generalized antiderivatives s ∈ L 2 , l o c ( ℝ + ) s\in L_{2,loc}(\mathbb{R}_{+}) . 1 Main results In the classical theory of the Sturm–Liouville operator defined by the differential expression l(y)=-y^{\prime\prime}+qy, where q\in L_{1,loc}(\mathbb{R}_{+}) , general results that do not require additional smoothness of the potential q are often formulated in terms of its antiderivative s , where q=s^{\prime} . In particular, the Molchanov criterion for the compactness of the resolvent [ 1 ] and its generalizations [ 2 , 3 , 4 , 5 ] characterize the growth of the antiderivative on segments of constant length. The nature of this phenomenon can be explained by the possibility of regularization of the differential equation l(y)=f for f\in L_{1,loc}(\mathbb{R}_{+}) : there is an equivalent system of equations, the coefficients of which are functions of the antiderivative: \begin{pmatrix}y(x)\\ y^{[1]}(x)\end{pmatrix}^{\prime}=\begin{pmatrix}s(x)&1\\ -s^{2}(x)&-s(x)\end{pmatrix}\begin{pmatrix}y(x)\\ y^{[1]}(x)\end{pmatrix}-\begin{pmatrix}0\\ f(x)\end{pmatrix}. The equivalent notation for the expression l(y) is following: l(y)=-(y^{[1]})^{\prime}-sy^{[1]}-s^{2}y,\ \mbox{where }y^{[1]}=y^{\prime}-sy. (1) It is expected that the general spectral properties of the regularized problem are expressed in terms of s . In their pioneering work [ 6 ] Savchuk and Shkalikov showed that regularization is an effective way of generalizing the Sturm–Liouville operator for potentials q\in W_{2,loc}^{-1}(\mathbb{R}_{+}) . Here W_{2,loc}^{-1}(\mathbb{R}_{+}) is the space of distributions with generalized antiderivatives s\in L_{2,loc}(\mathbb{R}_{+}) . The history of the issue and references are presented in [ 6 , 7 , 8 ] , on the regularization method and its development for high-order operators we recommend [ 9 ] . Here we consider an operator with a complex-valued potential, and our aim is to obtain generalizations of the known classical criteria for the compactness of the resolvent in terms of the antiderivative s of the potential q\in W_{2,loc}^{-1}(\mathbb{R}_{+}) . Throughout what follows q=s^{\prime} , and the derivative is understood in the sense of distribution theory. In a recent paper [ 5 ] we proposed a necessary condition for the compactness of resolvents for operators with arbitrary complex-valued potentials. q\in L_{1,loc}(\mathbb{R}_{+}) : \forall a>0\ \lim\limits_{x\to+\infty}\int\limits_{x}^{x+a}|q(\xi)|\,d\xi=+\infty. Of course, it is not relevant for q\not\in L_{1,loc}(\mathbb{R}_{+}) , and here we make up for this deficiency by proposing Theorem 1 — a stronger condition covering operators with arbitrary complex-valued q\in W_{2,loc}^{-1}(\mathbb{R}_{+}) . In [ 5 ] we found the operator L with essential spectrum, with potential q\in L_{1,loc}(\mathbb{R}_{+}) , \re q\geq 0 and |q|\to+\infty . The numerical range of L is contained in the right half-plane, and the resolvent is bounded in the left half-plane. It turned out that for a successful generalization of classical theorems of Molchanov type to the complex-valued case, one should require strict sectoriality for the minimal operator L_{0} — when the central angle of the sector with the numerical range of L_{0} is strictly less than \pi . For this reason, in the present paper we pay special attention to sectorial operators, for which we propose the corresponding theorems. In the special case of a real-valued potential of the form: q(x)=\sum_{x_{k}\in X}\alpha_{k}\delta(x-x_{k}), (2) and more general case, when the antiderivative s is a function of bounded variation, the issues of semi-boundedness and compactness of the resolvent have been studied previously. For example, Brasche [ 10 ] , arguing in the language of quadratic forms, proposed a sufficient condition for semi-boundedness of the quadratic forms: \int\limits_{\mathbb{R}}|y^{\prime}(x)|^{2}\,dx+\int\limits_{\mathbb{R}}|y(x)|^{2}\,ds(x),\quad y\in W_{2}^{1}(\mathbb{R})\cap L_{2}(\mathbb{R},s), where s is (neither necessarily finite nor necessarily positive) a Radon measure on \mathbb{R} . For the form to be semi-bounded, it is sufficient: \sup_{x>0}\bigl(s^{-}(x+1)-s^{-}(x)\bigr)\leq C, (3) where s=s^{+}-s^{-} is a decomposition into non-negative measures on \mathbb{R} . A generalization of Molchanov’s criterion for potentials of the form ( 2 ) was made in [ 11 ] . Proposing a criterion for compactness of the resolvent for semi-bounded operators, the authors require the Brachet condition ( 3 ). The classical result of Brink [ 2 ] suggests a possible weakening of this requirement to: \sup_{x>0}\bigl(s(x+a)-s(x)\bigr)\geq-C\ \mbox{for all }0<a\leq 1 In fact, the component s^{+} can compensate for an arbitrarily large s^{-} , ensuring semi-boundedness of the operator. This observation follows from theorem 4 below. In [ 12 ] the semi-boundedness of operators with potentials of the form ( 2 ) and discreteness of the spectrum were studied using the theory of Jacobi matrices. Note that the proposed sufficient condition for semi-boundedness [ 12 , Cor.5.26] follows from our theorem 4 . The present work is substantially inspired by the results of Ismagilov [ 4 ] and Zelenko [ 13 ] , and, of course, by the already mentioned work of Savchuk and Shkalikov [ 6 ] . Let s\in L_{2,loc}(\mathbb{R}_{+}) be a complex-valued function. For y\in AC_{loc}(\mathbb{R}_{+}) , we define y^{[1]}(x)=y^{\prime}(x)-s(x)y(x). For y^{[1]}\in AC_{loc}(\mathbb{R}_{+}) the following differential expression is correctly defined: l(y)=-(y^{[1]})^{\prime}-sy^{[1]}-s^{2}y. We introduce lineals: \displaystyle\mathscr{D}=\bigl\{y\in L_{2}(\mathbb{R}_{+})\,\bigl|\bigr.\,y,y^{[1]}\in AC_{loc}(\mathbb{R}_{+}),\,l(y)\in L_{2}(\mathbb{R}_{+})\bigr\}, \displaystyle\mathscr{D}_{0}=\bigl\{y\in\mathscr{D}\,\bigl|\bigr.\,y(0)=y^{[1]}(0)=0\,,\supp y\mbox{ --- is a compact}\bigr\}, and define operators in L_{2}(\mathbb{R}_{+}) : the minimal L_{0} on the domain \mathscr{D}_{0} and the maximal L_{M} on the domain \mathscr{D} , where: L_{0}y=l(y),\,y\in\mathscr{D}_{0},\quad L_{M}y=l(y),\,y\in\mathscr{D}. (4) Let us make a few reservations regarding the terminology used. Below s\in L_{2,loc}(\mathbb{R}_{+}) will be called the generalized antiderivative of the potential , or simply the antiderivative of the potential , meaning by the potential q=s^{\prime}\in W_{2,loc}^{-1}(\mathbb{R}_{+}) , where the derivative is understood in the sense of distribution theory. By \overline{l} we denote the differential expression constructed with \overline{s} — the complex conjugate function to s . By \overline{L}_{0} and \overline{L}_{M} we denote the minimal and maximal operators, defined by the differential expression \overline{l} . Their domains are, respectively, \overline{\mathscr{D}}_{0} and \overline{\mathscr{D}} . If the operator L is given, \widetilde{L} will denote its closure (if exists). The domain of \widetilde{L} is denoted as usual by \mathscr{D}(\widetilde{L}) . Everywhere \|\cdot\| is the norm in L_{2}(\mathbb{R}_{+}) . If I is an interval (open or closed, or semiinterval), \|\cdot\|_{I} is the norm in L_{2}(I) . The length of the interval I will be denoted by |I| . By C we denote arbitrary positive constants that may differ in adjacent formulas. Statement 1 . The domain \mathscr{D}_{0} is dense in L_{2}(\mathbb{R}_{+}) , the operator L_{0} is closable, its adjoint L_{0}^{*}=\overline{L}_{M} . The kernel \ker\widetilde{L_{0}}=\{0\} , the defect \widetilde{L_{0}} is finite: \codim\im L_{0}\leq 2 . Due to finite defect, the resolvents of all extensions of L_{0} (of all extensions that have resolvents) are either compact or not compact at the same time. Therefore, all results below are formulated in terms of the minimal operator. The following statement — a necessary condition for the compactness of the resolvent — is one of the key results of our study. Theorem 1 . If the minimal operator L_{0} has extensions with compact resolvents, then \forall a>0\quad\lim\limits_{x\to+\infty}\int\limits_{x}^{x+a}\int\limits_{x}^{x+a}|s(\xi)-s(\eta)|^{2}\,d\xi d\eta=+\infty. (5) Under additional requirements on s , the condition ( 5 ) is a criterion for the compactness of the resolvent. Before these results, we give some