Open AccessOptimization of information exchange in networks under conditions of uncertainty
Author(s) -
O. V. Senkov
Publication year - 2020
Publication title -
telekomunìkacìjnì ta ìnformacìjnì tehnologìï
Language(s) - English
Resource type - Journals
ISSN - 2412-4338
DOI - 10.31673/2412-4338.2020.042634
Subject(s) - information exchange , computer science , node (physics) , game theory , complete information , mathematical optimization , process (computing) , point (geometry) , value of information , relation (database) , control (management) , operations research , mathematics , mathematical economics , engineering , artificial intelligence , telecommunications , data mining , geometry , structural engineering , operating system
The article considers the problem of optimizing the exchange of information in the conditions of uncertainty that arises in the process of functioning of converged networks. The uncertainties formulated in the paper are characterized by discrepancies in the "interests" of different network nodes, varying degrees of awareness or incomplete data on the value of system parameters. Such uncertainties lead to a decentralized process management strategy in a converged network. The problem of interaction of network nodes is considered from the point of view of game theory, at the same time players act as separate nodes of system, the exchange of information between which is limited by conditions of uncertainty. Decision-making procedures are formulated for the following cases: when one node is the center and the other opposes it and the center knows information about the choice of the second player; when the center does not know the capabilities of the second player, and the uncertainty lies in the random nature of the strategy of the second player; network node synergy model. For the first case, the optimal control is based on the maximum action of the first station (center) and the maximum resistance of the other station (second player) and can use the maximum model. For the second case, the limitations of the strategy of the game (including the strategy of punishment of the player) are determined and it is proved that the gain, more than the maximum guaranteed result of the center, cannot be obtained on any other class of strategies. The relation of the center to the uncertainty in the model based on the synergetic concept of the operation of convergent networks is such that it seeks to maximize the linear convolution of the values of its criterion. The peculiarities of constructing a strategy for the exchange of information between individual nodes of the network in the presence and absence of isolated points in the set of choices of the player who opposes the center are studied. The developed model of interaction of heterogeneous subsystems of the telecommunication network in the conditions of uncertainty can be used for situations when nodes have incomplete information about both the system as whole and other nodes.