Searcharxiv⌕ Search

arXiv subjects

Dmytro R. Popovych

Publications and source records attributed to Dmytro R. Popovych.

9 recordsLinked to original sources
arXiv2025

Generalized symmetries of remarkable (1+2)-dimensional Fokker-Planck equation

Using an original method, we find the algebra of generalized symmetries of a remarkable (1+2)-dimensional ultraparabolic Fokker-Planck equation, which is also called the Kolmogorov equation and is singled out within the entire class of ultraparabolic linear second-order partial differential equations with three independent variables by its wonderful symmetry properties. It turns out that the essential subalgebra of this algebra, which consists of linear generalized symmetries, is generated by the recursion operators associated with the nilradical of the essential Lie invariance algebra of the Kolmogorov equation, and the Casimir operator of the Levi factor of the latter algebra unexpectedly arises in the consideration. We also establish an isomorphism between this algebra and the Lie algebra associated with the second Weyl algebra, which provides a dual perspective for studying their properties. After developing the theoretical background of finding exact solutions of homogeneous linear systems of differential equations using their linear generalized symmetries, we efficiently apply it to the Kolmogorov equation.

math-ph↗
arXiv2024

Algebraic method of group classification for semi-normalized classes of differential equations

We generalize the notion of semi-normalized classes of systems of differential equations, study properties of such classes and extend the algebraic method of group classification to them. In particular, we prove the important theorems on factoring out symmetry groups and invariance algebras of systems from semi-normalized classes and on splitting such groups and algebras within disjointedly semi-normalized classes. Nontrivial particular examples of classes that arise in real-world applications and showcase the relevance of the developed theory are provided. To convincingly illustrate the efficiency of the proposed method, we apply it to the group classification problem for the class of linear Schrödinger equations with complex-valued potentials and the general value of the space dimension. We compute the equivalence groupoid of the class by the direct method and thus show that this class is uniformly semi-normalized with respect to the linear superposition of solutions. This is why the group classification problem reduces to the classification of specific low-dimensional subalgebras of the associated equivalence algebra, which is completely realized for the case of space dimension two. Splitting into different classification cases is based on three integer parameters that are invariant with respect to equivalence transformations. We also single out those of the obtained results that are relevant to linear Schrödinger equations with real-valued potentials.

math-ph↗
arXiv2024

Improving physics-informed DeepONets with hard constraints

Current physics-informed (standard or deep operator) neural networks still rely on accurately learning the initial and/or boundary conditions of the system of differential equations they are solving. In contrast, standard numerical methods involve such conditions in computations without needing to learn them. In this study, we propose to improve current physics-informed deep learning strategies such that initial and/or boundary conditions do not need to be learned and are represented exactly in the predicted solution. Moreover, this method guarantees that when a deep operator network is applied multiple times to time-step a solution of an initial value problem, the resulting function is at least continuous.

cs.LG↗
arXiv2024

Generalized symmetries of Burgers equation

Despite the number of relevant considerations in the literature, the algebra of generalized symmetries of the Burgers equation has not been exhaustively described. We fill this gap, presenting a basis of this algebra in an explicit form and proving that the two well-known recursion operators of the Burgers equation and two seed generalized symmetries, which are evolution forms of its Lie symmetries, suffice to generate this algebra. The core of the proof is essentially simplified by using the original technique of choosing special coordinates in the associated jet space.

math.AP↗
arXiv2015

Canonical forms for matrices of Saletan contractions

We show that each Saletan (linear) contraction can be realized, up to change of bases of the initial and the target Lie algebras, by a matrix-function that is completely defined by a partition of the dimension of Fitting component of its value at the limit value of the contraction parameter. The codimension of the Fitting component and this partition constitute the signature of the Saletan contraction. We study Saletan contractions with Fitting component of maximal dimension and trivial one-part partition. All contractions of such kind in dimension three are completely classified.

math.RA↗
arXiv2014

Lie-orthogonal operators

Basic properties of Lie-orthogonal operators on a finite-dimensional Lie algebra are studied. In particular, the center, the radical and the components of the ascending central series prove to be invariant with respect to any Lie-orthogonal operator. Over an algebraically closed field of characteristic 0, only solvable Lie algebras with solvability degree not greater than two admit Lie-orthogonal operators whose all eigenvalues differ from 1 and -1. The main result of the paper is that Lie-orthogonal operators on a simple Lie algebra are exhausted by the trivial ones. This allows us to give the complete description of Lie-orthogonal operators for semi-simple and reductive algebras, as well as a preliminary description of Lie-orthogonal operators on Lie algebras with nontrivial Levi-Mal'tsev decomposition. The sets of Lie-orthogonal operators of some classes of Lie algebras (Heisenberg algebras, almost Abelian algebras, etc.) are directly computed. In particular, it appears that the group formed by the equivalence classes of Lie-orthogonal operators on a Heisenberg algebra is isomorphic to the standard symplectic group of an appropriate dimension.

math.RA↗
arXiv2014

Contractions with necessarily unbounded matrices

We prove that for each dimension not less than five there exists a contraction between solvable Lie algebras that can be realized only with matrices whose Euclidean norms necessarily approach infinity at the limit value of contraction parameter. Therefore, dimension five is the lowest dimension of Lie algebras between which contractions of the above kind exist.

math.RA↗
arXiv2010

Lowest dimensional example on non-universality of generalized Inönü-Wigner contractions

We prove that there exists just one pair of complex four-dimensional Lie algebras such that a well-defined contraction among them is not equivalent to a generalized IW-contraction (or to a one-parametric subgroup degeneration in conventional algebraic terms). Over the field of real numbers, this pair of algebras is split into two pairs with the same contracted algebra. The example we constructed demonstrates that even in the dimension four generalized IW-contractions are not sufficient for realizing all possible contractions, and this is the lowest dimension in which generalized IW-contractions are not universal. Moreover, this is also the first example of nonexistence of generalized IW-contraction for the case when the contracted algebra is not characteristically nilpotent and, therefore, admits nontrivial diagonal derivations. The lower bound (equal to three) of nonnegative integer parameter exponents which are sufficient to realize all generalized IW-contractions of four-dimensional Lie algebras is also found.

math-ph↗