Pavlova O. Agent-oriented information technology for assessing the initial stages of the software life cycle based on the ontological approach

The dissertation solves the current scientific and applied problem of automating the assessment of the initial stages of the software life cycle by developing agent-oriented information technology based on an ontological approach, which provides: automation of time-consuming, routine and error-prone task analysis requirements and its almost instantaneous implementation. The dissertation offers basic (universal) ontological models of non-functional characteristics-components of software quality, as well as ontological models of non-functional characteristics-components of quality of specific software, which are based on the requirements of ISO 25010: 2011 and ISO 25023: 2016 and provide a basis for the sufficient amount of information to assess the non-functional characteristics of the software. A model of intelligent agent activity based on ontological approach for evaluation of software requirements specifications has been developed, which is based on comparative analysis of ontologies and is a theoretical basis for developing methods of intellectual agent activity based on ontological approach. First time developed a method of intelligent agent based on an ontological approach to assess the initial stages of the software life cycle, which works on the basis of the developed model and evaluates the adequacy of information in the specification requirements to determine all non-functional characteristics of software. Provides a conclusion about the sufficiency or insufficiency of information in the specification, provides numerical estimates of the level of sufficiency of information to determine each non-functional characteristic of the software and to determine all non-functional characteristics-components of the software together, forms a list of attributes to supplement the specification of requirements to eliminate a person from the processes of information processing and acquisition of knowledge. Further elaborated: method of activity of the intellectual agent for the automated semantic analysis (parsing) of specifications of requirements to the software which carries out parsing of the specification, defines quantity and percent of the missing attributes, displays which attributes are missing for this or that subcharacteristic of the nonfunctional characteristic, forms a real ontology for non-functional characteristics that can be used by an intelligent agent based on an ontological approach to assess the initial stages of the software life cycle. First time developed: an agent-oriented information technology for assessing the initial stages of the software life cycle based on an ontological approach, which evaluates and improves the adequacy of information requirements to determine each non-functional characteristic separately and all non-functional characteristics together. This AOIT evaluates and provides an increase in the level of adequacy of information in the requirements specification to determine the non-functional characteristics of the software - the increase in the level of adequacy is from 4.71% to 27.79% (for example, from 58.23% to 86.02% for the requirements specification № 1 to the software agent for improving the security of computer systems software, from 81.26% to 85.97% for the specification №2, from 60.85% to 73.7% for the specification №3). The advantages of the developed AOIT are: automation of time-consuming, routine and error-prone task of analysis of requirements specifications and its almost instantaneous execution; hint where a rework on the requirements specification is needed (the user can view the missing attributes and see the areas of the specification that need extra attention, as well as which requirements need to be reworked); providing training for new specification developers, systems engineers, and project managers (using this AOIT helps them see the mistakes they may make and helps them recognize these mistakes in the work of others); assistance in developing high quality requirements; assistance in correcting and correcting requirements errors where they occur - in the early stages of the software life cycle - before they become more expensive to correct; providing a tool for selecting better specifications of software requirements; free internet access at any time without any registration. The economic effect of using the developed AOIT is the ability to save budget software projects for processing and correction (during the life cycle) of defects and errors that occur in the early stages of the life cycle - by demonstrating the weaknesses of software requirements that need to be finalized or revised at that time when they occur.