|
|
 Выпуск 36
- Morosanova N., Soloviev S. Formal properties of the Shortliffe scheme
We consider an isomorphic mapping of the Shortliffe scheme, which can significantly modify the functions of the scheme for calculating the certainty factors. We also investigate questions of construction of the isomorphisms generating schemes of plausible reasoning with desired properties.
- Pozdyayev V. Atomic optimization, part 1: search space transformation and one-dimensional problems
Optimization problems with polynomial objective function and inequality constraints are considered. A transformation of the moment theory-based solution method is presented, which allows to construct equivalent solution algorithms working in the augmented original search space instead of the moment space. One-dimensional problems are analyzed in detail.
- Agaev R. The region of convergence of the differential model of consensus
This paper is devoted to consensus problems in continuous multi-agent systems whose corresponding Kirchhoff matrix is diagonalizable and 0 is a simple eigenvalue of L. It is proved that the limiting matrix of the solution of the system of linear differential equations satisfying the initial condition is a eigenprojection of the Kirchhoff matrix L, which also determines and is defined the region of convergence to consensus of the DeGroot algorithm.
- Kondrashova E. Optimizing income function in controlled Markov queueing model
The controlled Markov queuing model M/M/n*/N* is studied. The income functional is being constructed on the trajectories of the controlled semi-markov process. The main problem is to optimize the income functional by adjusting the structure of the system.
- Chernykh N., Pakshin P. Algorithms for numerical solution of stochastic differental systems with switching diffusion
Mathematical models are considered of hybrid systems in the form of stochastic differential equations with Markovian switching of the diffusion component. An extension of Taylor schemes for numerical approximation of their solutions is proposed. Modeling results in SCILAB are presented to demonstrate efficiency of the obtained algorithms.
- Algazin G. Centralization and decentralization in basic game-theoretic models of organizational systems
The new interpretation of centralization and decentralization is given on the basis of a two-level model of an organizational system with unshared variables and asymmetric information. The classification of a wide range of basic models of the organizational systems is suggested and described. The classification is based on the degree of divergence of higher and lower layers’ goal priorities in their choice of shared variables.
- Goroshnikova T. Optimization model of the analysis of removal of the non-profitable enterprises from the company with algorithms of search optimum decisions
In this article we continue developingthe handy analytic technique of company restructuring for the sake of value maximization. The value of a company is assessed with the discounted cash flows technique. The underlying optimization models of mergers and acquisitions analysis and synthesis were published earlier in [4].
- Korgin N. Representing sequential resource allocation mechanism in form of strategy-proof mechanism of multi-criteria active expertise
We prove that any sequential resource allocation mechanism, which is strategy-proof for the domain of single-peaked preferences, may be represented as a strategy-proof mechanism of active expertise for the domain of multidimensional single-plateaued prefe¬rences – a generalized median voter scheme with a tie-breaking rule.
- Romanenko V. Optimizing technological process control of hub airport as queueing system with non-stationary flows and partial mutual assistance between channels
We set and solve the problems of optimal control program formation for a transportation service system in a hub airport treated as a queuing system with non-stationary flows and partial mutual assistance between channels. Policy recommendations are formulated for production units of hub airports.
- Salihov Z., Strizhko L., Novakovskaya A. Neural-network-based forecating model for controlof silver biosorption processes
We study the specific features of control in course of silver biosorption from biosorbents – secondary solutions originated from pharmaceutical industry production wastes. The plant state fore-casting model is implemented with a multi-layer multi-input neural network with inter-layer routing logic induced by logic of process-ing units switching. Existence of Koshi boundary conditions is assured at every switch point.
- Tuphanov I., Scherbatyuk A. Developing algorithms of auvs group behaviour for underwater local heterogeneities survey problem
A problem of search and size estimation of underwater local heterogeneities in a given area is considered. We introduce an approach to both localize heterogeneities and to estimate their dimensions and mass of substance dissolved. We suggest the method for planning work of a group of autonomous underwater vehicles /AUVs/, solving this problem. The corresponding discrete optimiza-tion problem is stated and two solution algorithms are considered. We also discuss a question of group actions re-planning in case of emergencies during an operation. The results of modeling for considered algorithms are provided.
- Asratian R., Lebedev V. Organization of protected http-interaction in multi-network environment
The problem is considered of information interchange in .NET-based geographically distributed information systems designed for complex multi-network environments consisting of several private enterprise networks. We suggest using the protected inter-network channels based on a system of proxy servers. The servers route HTTP-interactions using names of network nodes and resources.
- Pechnikov A., Chernobrovkin D. Adaptive crawler for external hyperlinks search and acquisition
We consider the web-crawler designed to find information about outgoing hyperlinks from a set of monothematic web-sites. Adaptive behavior of the crawler is formulated in terms of the multi-armed bandit problem. The experiments show that the choice of the adaptive algorithm of rational behavior depends on the subject of the considered set of web-sites.
|
|
