Assessing human performance influencing factors through LINMAP and Bayesian belief networks

Abstract

This study aims to identify and rank the Performance Influencing Factors (PIFs), which cause errors in human operations, by analyzing the failure weights and ranks of the tasks performed by every operator. Assessing these factors can mitigate human errors (HEs) and improve safety, efficiency, and job satisfaction. The Linear programming techniques for Multidimensional Analysis of Preference (LINMAP) and Bayesian Belief Networks (BBNs) were employed to analyze an aircraft tire manufacturing industry. In this method, all operators of workshops were evaluated. According to the data analysis, each operator’s tasks were weighted, and the potential error rate of each task was determined. PIFs for each workshop were then ranked and prioritized so that the most effective factors could easily be distinguished in order to identify the tasks where the operators had the highest rates of failure. The probability of HE was then obtained. In a predictive model, it is possible to determine when an error occurs and which factors are the most effective in its occurrence. This paper proposes an approach to the easy, inexpensive, and rapid classification of PIFs by determining their correlations through conditional possibilities. The proposed approach is capable of classifying not only PIFs but also the PIF-related tasks with the greatest effects.