Lishchuk K. Linear models and methods for automated support in the hierarchical decision-making systems

Dissertation is dedicated to solving actual scientific problem of automated support of decision-making in hierarchical problems criterion choice. The proposed and justified criteria and constructive mathematical models of linear and convex quadratic programming for finding weights of objects of empirical matrix of pairwise comparisons. The proposed two modifications of AHP, which differ from the other possibility of solving problems of hierarchical multi selection of high dimensionality. Shown that in solving the problem of multi selection through AHP weight alternatives to lower-level criteria are defined using linear convolution, which eliminates the pairwise comparison of alternatives on the lower level hierarchical tree. The mathematical apparatus operative correction of assessments of one of the alternatives that ensures the selection of this alternative is based on the use of replacement ratio. Grounded basic strategies of the person who decides. The algorithms of operative correction of decision-making tasks in hierarchical multi choice are proposed. Based on the proposed mathematical apparatus developed information decision support system, which is used to solve practical proplemes of arbitrary dimension selection criterion by AHP. Key words: Analytic Hierarchy Process, the ratio of substitution, surgical correction.