Hybrid knowledge representation model for a humanitarian response system IT project

The object of research is the process of managing IT projects for the development of humanitarian response systems. The features of knowledge representation in the process of IT project management are considered. The requirements for a hybrid model of knowledge representation are formulated, such as the use of previous experience, multilevel, interoperability and consideration of uncertainty. A multidimensional ontological model is considered, which contains five levels: fundamental, core, upper domain, lower domain and operational. The peculiarity of the model is the combination of three ontologies at the core level: IT project management, humanitarian response system and humanitarian response processes. The combination of case-based reasoning with a multi-level ontology using fuzzy associative relations is used as the basis for knowledge representation. To take into account uncertainty the model is extended with fuzzy elements and fuzzy inference procedures. The model inference process includes the stages of forming a model of the current situation, searching for the most relevant case, mapping case parameters to the ontology, selecting a fragment of the ontology to solve the problem, obtaining uncertain parameters through fuzzy inference, enriching the selected fragment of the ontology, mapping the enriched fragment of the ontology to a solution, adapting the resulting solution, and storing a new case. The experimental study was conducted on the developed ontology using cases created based on the analysis of completed IT projects. The ontology was supplemented with fuzzy elements and fuzzy inference procedures. The experiment showed that the quality of classification of a situation using the hybrid model under conditions of uncertainty of some parameters is 16 % better than when using classical case-based reasoning.