Open Repository of Keldysh Institute of Applied Mathematics of RAS
Not a member yet
5498 research outputs found
Sort by
Intelligent core decision support system
No full textAbstract:
The paper discusses conceptual issues of the organization of information systems focused on decision support tasks. An invariant with respect to the subject domain decomposition of a decision support system into a number of subsystems is proposed. The conceptual scheme of information flows allowing to create systems of distributed artificial intelligence is created. GRID technologies which are one of the perspective directions of development of the software of the multi-criteria analysis of alternatives and optimization are investigated.Note:
Research direction:Mathematical modelling in actual problems of science and technic
Solution of stiff Cauchy problems with explicit schemes with geometrical-adaptive step selection
No full textAbstract:
We propose an explicit numerical method for solution of stiff Cauchy problems. The method implies explicit schemes and step selection procedure based on curvature of the integral curve. We propose explicit formulae for the curvature. For the Runge-Kutta schemes with up to 4 stages, the sets of the scheme coefficients are provided. Verification of the method is performed on a test problem with a known exact solution. We show that the method possesses the same accuracy and robustness as implicit methods and sufficiently excels them in efficiency.Note:
Research direction:Mathematical problems and theory of numerical method
Numerical simulation of low-speed flows around of power plant using NOISEtte
No full textAbstract:
This work presents some results of numerical simulation of flow around the power plant with vertical axis of rotation based on the research code NOISEtte. The low Mach preconditioning approach is used to numerical solution of the system of compressible viscous equations and this NOISEtte option provides a set of multivariate calculations.Note:
Research direction:Mathematical modelling in actual problems of science and technic
Searching for centrally symmetric generating solutions in Hill problem
No full textAbstract:
The non-integrable classical Hill problem is embedded into more general one which links it with integrable Kepler problem in uniformly rotating frame. It makes possible to apply the method of regular normal form to the last one and to obtain so called generating solutions with different type of symmetries. New class of generating solutions with central symmetry is obtained. The corrections to the initial conditions and period of such solution are computed for continuation of such solutions up to periodic orbits of the planar circular Hill problem.Note:
Research direction:Mathematical modelling in actual problems of science and technic
Solving the overdetermined problems for systems of linear ordinary differential equations
No full textAbstract:
The ways of applying the least squares method for solving the overdetermined systems of linear ordinary differential equations with possibly redundant boundary conditions are considered. Such problems generally have no solutions.
The methods proposed for solving these problems take into account both the equations themselves and the corresponding boundary conditions. In one of these methods the order of equations of the considered system is increased. In this case additional boundary conditions are formulated for the system. The model examples of application of the methods and their comparison are given.Note:
Research direction:Mathematical problems and theory of numerical method
On temporal stability of the Poiseuille flow in a channel of elliptic cross-section
No full textAbstract:
A numerical model for the temporal stability analysis of the Poiseuille flow in a channel of constant elliptic cross-section is described. In particular, an algorithm for computing the maximum possible amplification of the average kinetic energy density of disturbances with the use of a spectral reduction is described. This reduction allows us to significantly reduce the computational costs as well as to eliminate artefact disturbances appearing because of the approximation errors. The maximum amplification of the average kinetic energy density of disturbances is computed for channels with round and elliptic cross-sections. It is shown that its absolute maximum in the case of the round and elliptic channel cross-sections is attained on disturbances which possess different symmetries.Note:
Research direction:Theoretical and applied problems of mechanic
Comparison of two direct interplanetary trajectory optimization techniques
No full textAbstract:
In this work, two direct methods of low-thrust trajectory optimization are implemented and compared. The first method is based on optimization of the thrust acceleration, whereas the second method optimizes the impulses that approximate the thrust arcs. The methods are applied to the sample problem of a transfer from the Earth to Mars. The results are obtained on a personal computer in MATLAB and on a multiprocessing computing system.Note:
Research direction:Theoretical and applied problems of mechanic
Zero-mean interpolation inequality on the sphere
No full textAbstract:
We prove multiplicative interpolation inequalities for the imbeddings of the Sobolev space H1(S2) into Lq for q ∈ [1,∞). The case of zero-mean functions is considered and similar inequalities for tangent vector functions. The corresponding constants are explicitly found with sharp rate of growth with respect to q as q → ∞. In the one-dimensional periodic case the corresponding inequalities are proved in the critical case H 1/2 (S1) →Lq(S1) for all q ∈ [1,∞).Note:
Research direction:Mathematical problems and theory of numerical method
Molecular dynamic modeling of thermophysical properties of copper in the region of the melting point
No full textAbstract:
Method of molecular dynamics is used for calculation of thermodynamic properties of copper: temperature dependence of melting heat and temperature and pressure dependence of specific heat, coefficient of linear expansion and density. The obtained dependences are compared with experiment. These dependences can be used as input data for the continuum model of pulsed laser heating of matter.Note:
Research direction:Mathematical modelling in actual problems of science and technic
Development of exponential integrator based on classical Runge—Kutta method and it's application for solving stiff systems
No full textAbstract:
The paper deals with numerical method for solving stiff systems of ordinary differential equations. The new method RK4exp is derived on the base of classical four-stage Runge—Kutta method. The convergence of RK4exp is proved and it's stability function is calculated. Theoretical results are verified with numerical experiments.Note:
Research direction:Mathematical problems and theory of numerical method