« ACASYA »: a knowledge-based system for aid in the storage, classification, assessment and generation of accident scenarios. Application to the safety of rail transport systems

Various researches in artificial intelligence are conducted to understand the transfer of expertise problem. Today we perceive two major independent research activities: the acquisition ofknowledge which aims to define methods inspired specially from software engineering and cognitive psychology to betterunderstand the transfer of expertise, and the automatic learning proposing the implementation of inductive, deductive, abductivetechniques or by analogy to equip the system of learning abilities. The development of a knowledge-based support system ACASYA for the analysis of the safety guided transportsystems insisted us to use jointly and complementary both approaches. The purpose of this tool is to first, to evaluate the completenessand consistency of accidents scenarios and secondly, to contribute to the generation of new scenarios that could helpexperts to conclude on the safe character of a new system. ACASYA consists of three learning modules: CLASCA,EVALSCA and GENESCA dedicated respectively to the classification, evaluation and generation accident scenarios

Keywords:

Transport system , Safety , Accident scenario , Acquisition , Assessment , Artificial intelligence , Expert system , Machine learning