К вопросу об автоматизации процедур определения победителей конкурсов научно-технических проектов и стартапов
The process of selecting a contest winner is presented in the form of a multistage decision-making procedure. At the first stage, it is necessary to generate a list of alternative projects (applications for participation in a particular contest). At the second stage a list of criteria to select the alternatives is formed. Next, it is required to determine numerical or linguistic estimation scales based on the selected criteria. It is assumed that before the contest is held several ordered gradations are determined for each criterion and estimation of each of the proposed alternative projects can be attributed to one of them. At the next stage, the problem of determining the relative importance of criteria and corresponding gradations for these criteria is solved. Then comes the procedure itself, which compares the received applications using the selected criteria and estimation scales. The last task is very labour intensive and its automation will allow to simplify greatly the work of the juries.
This paper examines the questions related to building automated systems aimed to determine the winners of scientific and technical project competitions based on the decision-making support methods. It offers an integrated approach including technologies and procedures of analytical hierarchies methods (can be used for the purpose of gradation weight on the criteria scales), group methods of expert estimation of multicriteria alternatives – the method of preferences or rank method (can be used for purposing of scales gradation on scales of criteria and for the purpose of criteria weight), methods of group decisions coordination using the preference functions or decision-maker relations (it is used for coordination of expert numerical estimates of gradation on scales of criteria). The examples of calculations illustrating this approach are given.