Kuzmina N. Information technology of group decision making support in distributed system

The dissertation work is dedicated to increase the efficiency of the distributed group decision support systems. A new information model of a distributed group decision support system is proposed, which allows to combine all elements of a distributed system and provides an opportunity to influence the initial probabilities of alternative decision-making by introducing current user preferences. A method of preparing data for group decision-making using a Bayesian network is proposed, which, unlike the existing ones, allows to form alternatives for user selection by statistical analysis of the training sample, which allows to increase group decision-making efficiency by offering the most probable initial alternatives for user selection. A dynamic Bayesian network for the formation of user choice alternatives is proposed, which uses apriori probabilities of alternatives and is dynamically updated by users during each decision-making stage, which allows to propose the most probable alternatives at each stage of user decision-making process until all the alternatives are chosen. Algorithmic support of the distributed system of group decision-making support is developed. Developed and programmatically implemented group decision-making technologies that take into account the preferences of each user. Software and methodological support for group programming training has been developed, which uses the proposed information technology with modules of program codes parcing.