Leshchynska O. Intelligent synthesis methods of programs semantic models

The work is devoted to development of intelligent synthesis and verification methods of programs semantic models. Such methods are able to accelerate the software development process. Program semantic model and the structure of semantic annotation for model elements, that formalizes the essence of annotated element, its behavioral requirements and context-sensitive usage guidelines, are introduced in thesis. The structure of ontology has been developed that includes concepts and relations for semantic model elements representation, their attributes, semantic annotation elements, general concepts taxonomy to fill annotation instances slots. On the basis of program semantic models and ontological knowledge base such intelligent methods have been developed: semantic model synthesis via refined components matching by case-based reasoning in ontological knowledge model; method of case-based reasoning system reinforcement learning. Verification of program model satisfiability is refined to consider behavioral elements. Method of semantic-syntactic analysis was refined to use multicomponent formal language's model, that unify analysis of textual and graph languages, integrate stages of syntactic and semantic analysis. Models and methods are implemented as prototype of programs semantic models synthesis system.