Kadenko S. Experience-based decision-making support methods using ordinal estimates

The thesis research object is the expert estimate based decision-making process. The thesis goal is to improve the expert estimation-based decision-making process through expanding the mathematical apparatus used for expertise conduction. The research is based on the hierarchical approach to decision-making. The mathematical means of relative criterion weight coefficient calculation were developed based on iterative neuronal network parameter fine-tuning algorithms and linear programming apparatus. The key thesis results are the following ones: relevance of using multi-criteria ordinal estimates of alternatives for criterion weights calculation is proved; two criterion weight calculation methods are developed; a decision-making support (DSS) structure modification allowing to process expert information in several ways is suggested. The suggested mathematical apparatus is implemented in an automated expert workplace model. The thesis' scientific novelty lies, first of all, in the fact that it's ordinal and not cardinal estimation experience that is used for criteron weights calculation. The thesis results were used while working on budgetary themes of the Institute for information recording. Suggested methods are being implemented into "Solon" DSS. Materials of the thesis were also used in scientific research work on the theme "Development of the scientific background for higher educational establishment rating using the world experience". The work was done on request from the Ministry of education and science of Ukraine in the framework of a tender contest held by the State fundumental research fund. The author's research results relating to this specific research are to be used in the higher educational establishment rating system, currently under development.