Mehmet Gromchuk - Department of Statistics, Faculty of Science, University of Istanbul, Istanbul, Turkey

Production systems confront many accidents, malfunctions and quality defects due to human errors. Human reliability analysis (HRA) has been developed to identify and quantify the human error probability. Each HRA method has its advantages and disadvantages and has proposed for a specific environment or system and utilizes some unreal assumptions to make the problem easier to solve this assumption derive the HRA method from reality and the obtained results are unreliable. To overcome this issue. We propose an Artificial Intelligence System (AIS) in cooperation with Response Surface Method (RSM) to provide a new HRA method (ARHRA) and make HRA closer to reality. This method proposes a framework to calculate the effects of performance shaping factors on human error probability (HEP) with AIS and RSM. The proposed model has been applied to a real case and the provided results show that human reliability can be calculated more effectively using ARHRA method. 

Realiability, human resource, Error, response surface method