Institute de Cercetare / Research Institutes

Permanent URI for this communityhttps://msuir.usm.md/handle/123456789/13367

Browse

Start date: 2022Has files: Yes

Search Results

Now showing 1 - 10 of 816
  • No Thumbnail Available
    Item
    PERSIST: A New Probabilistic Model For Data Denoising [Articol]
    (2024) Scrob, Sergiu
    The paper proposes a more efficient solution for reducing the data noise level, using probability estimation for radius-based spatial inference and sampling techniques.
  • No Thumbnail Available
    Item
    Informational Extended Games And Their Applications [Articol]
    (2024) Novac, Ludmila
    In this article, we analyse informational extended games, i.e., games in which the players choose their actions simultaneously, with assumption that they have some information about the future strategies which will be chosen by other players. For all informational extended games of this type we assume that players’ payoff functions are common knowledge. Under these assumptions we define the noncooperative informational extended games and analyse Nash equilibrium. As a particular case of the non-cooperative informational extended games, we analyse the class of bimatrix informational extended games and we present an example of game in order to show the possibility to use the information for this class of games.
  • No Thumbnail Available
    Item
    Strongly Polynomial Primal-Dual Algorithms for Concave Cost Combinatorial Optimization Problems [Articol]
    (2024) Magnanti, Thomas L.; Stratila, Dan
    We introduce an algorithm design technique for a class of combinatorial optimization problems with concave costs. This technique yields a strongly polynomial primal-dual algorithm for a concave cost problem whenever such an algorithm exists for the fixed-charge counterpart of the problem. Using our technique, we obtain a 1.61-approximation algorithm for the concave cost facility location problem, an exact algorithm for the concave cost lot-sizing problem, and a 4-approximation algorithm for the joint replenishment problem with general concave individual ordering costs.
  • No Thumbnail Available
    Item
    Homogeneous Linear Recurrent Games [Articol]
    (2024) Lazari, Alexandru
    Games defined on deterministic systems with homogeneous linear recurrent dynamic are studied. It is proved that they represent homogeneous linear recurrent systems and an efficient method for determining the generating vector is proposed. This enables the usage of the advanced properties of homogeneous linear recurrences for their algebraic characterization.
  • No Thumbnail Available
    Item
    Characteristic Function of the Markov Random Flight in Higher Dimensions [Articol]
    (2024) Kolesnik, Alexander D.
    Two series representations of the characteristic function of the multidimensional symmetric Markov random flight, are presented. These series are the decompositions of the characteristic function with respect to Bessel functions and with respect to time variable, whose coefficients are given by recurrent relations, as well as in the form of special determinants. Basing on these series representations, an asymptotic formula for the second moment function of the process, is obtained.
  • No Thumbnail Available
    Item
    Some Analytical Properties of the Laplace Transform of the Characteristic Function of the Three-Dimensional Markov Random [Articol]
    (2024) Kolesnik, Alexander D.
    The analytical properties of the Laplace transform of the characteristic function of the three-dimensional symmetric Markov random flight, are studied. The only singular point (ordinary pole) and the residue at this point are evaluated, which gives the first coefficient (by the negative power of complex variable) in the Laurent decomposition of the Laplace-Fourier transform of the transition density of the process.
  • No Thumbnail Available
    Item
    Data Parallelization for Solving Bimatrix Games [Articol]
    (2024) Hâncu, Boris; Turcanu, Calin
    Contemporary decision-making problems are very complex and require the processing of a very large volume of data. Thus, for the mathematical modelling of these processes, it is necessary to take into account the big data problems. The data is too big to be stored and processed by a single machine. In many large-scale solutions, data is divided into partitions that can be managed and accessed separately. In order to solve such problems in real time, parallel algorithms are built and then implemented on various types of parallel computing systems. In this paper, we will analyze the ways to build parallel algorithms, and especially, data parallelization, for a class of noncooperative games, bimatrix games.
  • Thumbnail Image
    Item
    Transportation Problems, the Braess Paradox and Network Coordination [Articol]
    (2024) Gorbachuk, Vasyl; Dunaievskyi, Maksym; Havrylenko, Serhiy
    Traveling through the transport network or sending information packets via the Internet is implicitly based on game-theoretic considerations: a specific decision-maker (DM), choosing his or her route, takes into account the probability of congestion depending on all DMs, that is, other routes. Based on similar considerations, it is possible to develop models for network traffic. These models explain some paradoxical observations where increasing the capacity of a given network can slow down its traffic under certain circumstances.
  • Thumbnail Image
    Item
    Algorithm to Minimize the Worst-Case Regret Function [Articol]
    (2024) Godonoaga, Anatol; Chumakov, Borys
    This paper examines mathematical model of decision-making under conditions of uncertainty, when Savage’s regret function is used as the objective function. Numerical algorithm for solving the corresponding problems are proposed, based on the implementation of parallel minimization of indicators for each state of nature and the indicator of greatest regret. The algorithms are developed based on the generalized gradient projection method.
  • Thumbnail Image
    Item
    Strong Edge-Colorings of Some Lexicographic Products of Graphs [Articol]
    (2024) Drambyan, Aram
    The edge-coloring of a graph G is strong if edges at distance 0 or 1 receive different colors. The minimum number of colors required for a strong edge-coloring of G is called strong chromatic index of G and denoted by χ′ s(G). In this paper, we determine the exact value of the strong chromatic index of lexicographic products of graphs G and H when Δ(G) ≤ 2 and H is an arbitrary graph, and provide an upper bound on χ′ s(G[H]) when Δ(G) = 3 and H is an arbitrary graph.