Vol 523, No 1 (2025)

For the integral Radon transform, classical formulas for inversion of the integrand function are known, provided that it is smooth. However, this restriction does not fully correspond to the application of the results in sounding theory, which is the main area of application of the Radon transform. It would be more natural to assume that discontinuities of the first kind are admissible for integrands. The paper presents a number of inversion formulas proved by the authors for piecewise continuous functions. A comparison of the obtained formula variants is carried out and preliminary recommendations on their use for numerical algorithms are given.

The work is devoted to basic categorial grammars with unique type assignment (BCGUTA). For this class, a number of algorithmic properties are examined. It is proven that, for an arbitrary context-free language L, the problem of verifying whether is generated by a grammar from the BCGUTA class is algorithmically undecidable. Furthermore, it is proven that for any two BCGUTA grammars, the problem of determining the emptiness of the intersection of the languages generated by these grammars is also algorithmically undecidable.

The problem of interpolation in the mean of a function on known integrally averaged values by an integro quadratic spline is considered. It is shown that the integro quadratic spline can be defined via the interpolation cubic spline. Since the interpolation cubic spline is studied quite well, well-known error bounds for interpolation and some of its properties can be transferred to the integro quadratic spline. The points of superconvergence of the integro spline are found, i.e. the points at which the spline or its derivatives have a higher order of approximation.

The three-dimensional exterior Neumann problem for the Helmholtz equation is considered. Using the potential method, it is reduced to a boundary weakly singular Fredholm integral equation of the second kind, which is solved numerically. The accuracy is increased and the computational complexity of the numerical solution algorithm is reduced by averaging the kernel of the integral operator and localizing its singular part during discretization using simple analytical expressions. Examples of using this approach in the numerical solution of the original problem are given.

A mixed problem for the B-hyperbolic equation in Euclidean domains with different locations relative to singular coordinate hyperplanes is considered. In each of these domains, energy integrals with respect to the Lebesgue—Kipriyanov integral measure with weak and strong singularities are introduced. The absence of energy flow through coordinate singular hyperplanes, which are the internal boundary of mirror-symmetric regions in Euclidean space, is proven. If solutions to these problems exist, their uniqueness is proven.

Consider E(G, k) — the set of all sizes (numbers of edges) of induced subgraphs of size k in a given graph G with n vertices. For the binomial random graph G = G(n, p), we proved that, for each α > 0 and ε small enough, the set E(G, k) with high probability contains a large segment for all k such that (ln n)1+α < k < εn. We also found the asymptotic length of this segment.