


Vol 55, No 3 (2019)
- Year: 2019
- Articles: 15
- URL: https://journals.rcsi.science/0012-2661/issue/view/9363
Ordinary Differential Equation



Information Meaning of Entropy of Nonergodic Measures
Abstract
The limit frequency properties of trajectories of the simplest dynamical system generated by the left shift on the space of sequences of letters from a finite alphabet are studied. The following modification of the Shannon-McMillan-Breiman theorem is proved: for any invariant (not necessarily ergodic) probability measure μ on the sequence space, the logarithm of the cardinality of the set of all μ-typical sequences of length n is equivalent to nh(μ), where h(μ) is the entropy of the measure μ. Here a typical finite sequence of letters is understood as a sequence such that the empirical measure generated by it is close to μ (in the weak topology).



Strong Embeddability of Time-Invariant Nonlinear Differential Systems in Linear Differential Systems
Abstract
We consider a pair (Σ, w) consisting of a time-invariant nonlinear differential system Σ and a real piecewise smooth function w defined on the state space of the differential system. The notion of strong embeddability of the pair (Σ, w) in a linear differential system is introduced. A criterion for the finite-dimensional-strong embeddability of the pair is obtained, linear embedding systems are described, and the explicit representation of first integrals is given for systems in finite-dimensional-strongly embeddable pairs. The class of first integrals determining the general solution of a finite-dimensional-strongly embeddable pair is singled out.






Dulac-Cherkas Test for Determining the Exact Number of Limit Cycles of Autonomous Systems on the Cylinder
Abstract
For real autonomous two-dimensional systems of differential equations with continuously differentiable right-hand sides 2π-periodic in one of the variables, we consider the problem of determining the exact number of limit cycles of the second kind on the cylinder. If the system has no equilibria, then we propose to solve this problem by two methods based on a successive two-step application of the Dulac-Cherkas test or one of its modifications, which permits determining closed transversal curves dividing the cylinder into subdomains surrounding it and such that the system has precisely one limit cycle of the second kind in each of them. The efficiency of the proposed methods is illustrated with examples of systems corresponding to the Abel equation, for which the number of limit cycles on the whole phase cylinder is established.



Solutions of the Fourth-Order Equation in the Generalized Hierarchy of the Second Painlevé Equation
Abstract
We consider the analytic properties of solutions of the fourth-order higher analog of the second Painlevé equation and study the local properties of solutions, Bäcklund transformations, rational solutions, and their representation via generalized Yablonskii-Vorob’ev polynomials.



Extremum Condition and Stability Tests for Solutions of Gradient Systems
Abstract
We study the Lyapunov stability of equilibria of gradient systems. We describe the class of functions generating the right-hand side of a gradient system for which sufficient condition for a nonstrict local minimum are also stability conditions for the equilibria. The corresponding extremum conditions for functions of several variables are given. Stability tests for completely solvable systems with a multidimensional independent variable are stated.



Partial Differential Equation
Cauchy Problem and the Second Mixed Problem for Parabolic Equations with a Dirac Potential Concentrated at Finitely Many Given Points
Abstract
We prove the existence and uniqueness of a classical solution of the Cauchy problem and the second mixed problem for parabolic equations whose potential is a linear combination of values of this solution at finitely many prescribed points.



Classical Solution of a Problem with Integral Conditions of the Second Kind for the One-Dimensional Wave Equation
Abstract
for the one-dimensional wave equation given in a half-strip, we consider a boundary value problem with the Cauchy conditions and two nonlocal integral conditions. Each integral condition is a linear combination of a linear Fredholm integral operator along the lateral side of the half-strip applied to the solution and the values of the solution and of a given function on the corresponding base of the half-strip. Under the assumption of appropriate smoothness of the right-hand side of the equation and the initial data, we obtain a necessary and sufficient condition for the existence and uniqueness of a classical solution of this problem and propose a method for finding it in analytical form. A classical solution is understood as a function that is defined everywhere in the closure of the domain where the equation is considered and has all classical derivatives occurring in the equation and in the conditions of the problem.



Control Theory
State Estimation for Linear Time-Varying Observation Systems
Abstract
We consider the state observation and estimation problems for linear time-varying systems of ordinary differential equations. We show that the uniform observability is equivalent to the approximate observability, i.e., the possibility of using delta sequences to estimate the current state with arbitrary accuracy without differentiating the output function. We propose a method for constructing state estimators for linear time-varying systems based on the quasi-differentiability of the coefficients with respect to a specially constructed lower-triangular matrix. For uniformly observable systems with quasi-differentiable coefficients, we obtain conditions for the existence of an exponential observer and describe a constructive method for designing such observers.



Necessary Conditions for Optimality in Problems of Optimal Control of Systems with Discontinuous Right-Hand Side
Abstract
We consider the problems of optimal control of a dynamical system whose right-hand side is discontinuous in the state variable and is linear in the control with sufficiently smooth coefficients in each of the half-spaces into which the space is divided by the switching hyperplane. The main attention is paid to the situation where there exist intervals on which the optimal trajectory lies on the switching surface. New nondegenerate necessary conditions for optimality are stated and proved in the maximum principle form. The obtained optimality conditions are compared with the already known conditions.



Construction of Observers for a Delay Differential System with One-Dimensional Output
Abstract
We construct systems for estimating the solutions of a spectrally observable linear autonomous system with commensurable delays from the one-dimensional output with various dynamic properties: (i) the estimation error is the output of an asymptotically stable system of the same type (an asymptotic observer); (ii) starting from some time, the output of the observer system is identically equal to the solution of the original system (a finite observer), and simultaneously, the observer is asymptotic (a complete observer). The results are illustrated with examples.



Synthesis of Observers for Linear Systems of Neutral Type
Abstract
Two types of observers are proposed for linear autonomous differential-difference systems of neutral type. The first observer can be applied to completely identifiable systems (generalization of the spectral observability property to systems of neutral type), and it asymptotically exactly reconstructs the current state of the system. The second observer can be applied to systems without the above-cited property. The estimation error of the second observer is described, and conditions under which it asymptotically exactly reconstructs the current state of the system are proposed, as well as conditions under which the estimation error is bounded. The results are illustrated with examples.



Numerical Method
Multipoint Boundary Value Problem for the Lyapunov Equation in the Case of Weak Degeneration of the Boundary Conditions
Abstract
Sufficient coefficient conditions for the unique solvability of a multipoint boundary value problem for a matrix Lyapunov differential equation are obtained in the case of weak degeneration of the boundary conditions. An iteration solution algorithm based on the computational scheme of the classical successive approximation method is proposed.



Monotone Finite-Difference Schemes of Second-Order Accuracy for Quasilinear Parabolic Equations with Mixed Derivatives
Abstract
We consider the initial-boundary value problem for quasilinear parabolic equation with mixed derivatives and an unbounded nonlinearity. We construct unconditionally monotone and conservative finite-difference schemes of the second-order accuracy for arbitrary sign alternating coefficients of the equation. For the finite-difference solution, we obtain a two-sided estimate completely consistent with similar estimates for the solution of the differential problem, and also obtain an important a priori estimate in the uniform C-norm. These estimates are used to prove the convergence of finite-difference schemes in the grid L2-norm. All theoretical results are obtained under the assumption that some conditions imposed only on the input data of the differential problem are satisfied.


