Elizaveta V. Podobedova, Marina S. Zhukova, Leonid V. Arshinskiy

Irkutsk state transport university

The paper considers an approach to the development of an expert system for the analysis of accidents in railway transport. The stages and features of the development of its knowledge base, the possibilities of a tool environment and other issues are discussed. The prototype of the system was developed only for the analysis of derailments, but the knowledge model adopted in it makes it relatively easy to expand the system to other couses of accidents. The system is implemented in an environment using inference based on logic with vector semantics, which allows you to work with both reliable and fuzzy, uncertain and contradictory information.

expert system, vector semantics, accident investigation, railway transport