Ivanilov A. Relational construction tools of logic networks models and their application in an artificial intellect

The dissertation is devoted working out of logic networks models formal construction tools. These means are based on theorems from normalization section of the relational data-bases theory and allow to transform any predicates algebra equations to logic networks models. Logic networks hardware realization on programmed logic integrated schemes forms the new high-efficiency processors for processing of a natural language models. For the first time the method of the relational algebra operations formal description in final pre-dicates algebra language is developed. Has received the fur-ther development of predicates decomposition method which is based on statements about functional, multiple-valued dependences and dependence of connection of relational relations, Hez and Feygin theorems and statements about predicate dependences. For the first time the method of functional predicates binary decomposition is developed and proved. The model of a logic network for verbal word forms inflectional processing is constructed. The estimation of the received model size is executed and results of hardware modelling of a logic network on integrated schemes of pro-grammed logic are presented.