Hilbert spaces of solutions to polynomial Dirac equations, Fourier transforms and reproducing kernel functions for cylindrical domains (Q2492072): Difference between revisions

Let \(\mathbf{D}_\mathbf{z}:=\sum_{j=1}^k\partial_{z_j}e_j\) be the Euclidean Dirac operator in \(\mathbb R^k\); functions defined in an open subset of \(\mathbb R^k\) and satisfying there \(\mathbf D_\mathbf z f=0\) are called left monogenic with respect to the vector variable \(\mathbf z:=\sum_{j=1}^kz_je_j,\;z_j\in\mathbb R\). The authors are interested in the solutions of the following equation: \((\mathbf D_\mathbf z-\lambda_1)\cdots (\mathbf D_\mathbf z-\lambda_p)f(\mathbf z)=0\), where \(\lambda_1,\dots,\lambda_p\) are arbitrary non-zero complex numbers; the equation itself is called polynomial Dirac equation. Section 1, Introduction and Basic Notions, gives a brief survey of the study of Bergman-type and Hardy-type spaces in the framework of Clifford analysis. Section 2, Reproducing kernels for the unit ball for polynomial Dirac equations, begins with the following fact: the functions from the Bergman-type space for polynomial Dirac equation satisfy the Bergman condition, i.e., the evaluation operator is bounded. Of course, this ensures immediately the existence of the reproducing kernel function. It's appeared that in the context of Hardy-type spaces for polynomial Dirac equations an analogy can be established for the cases \(p=1\) and \(p=2\), but whenever \(p>2\) the corresponding spaces do not possess reproducing kernel functions. Finally in the Section, there are given explicit representation formulas for the Bergman kernel in a case of arbitrary complex polynomials that split in pairwise different linear factors, as well as for the Szegö kernel of the Hardy spaces with \(p=1\) and with \(p=2\), \(\lambda_1\neq\lambda_2\). In particular, for \(p=2\) and \(\lambda_1=-\lambda_2\) this leads to explicit formulas for the Bergman and the Szegö kernels to the Helmholtz equation \(\Delta f+\lambda^2f=0\). In the concluding Section 3, Reproducing kernels for cylinders, the explicit formulas are obtained for the unit infinite cylinder, for a half-infinite cylinder and for bounded finite cylinders.