A novel evidence theory model dealing with correlated variables and the corresponding structural reliability analysis method

Evidence theory serves as a powerful tool to deal with epistemic uncertainty which widely exists in the design stages of many complex engineering systems or products. However, the traditional evidence theory model cannot handle parameter correlations that may have profound influences on the reliability analysis results. This paper is supposed to develop a novel evidence theory model with consideration of parameter correlations and its corresponding structural reliability analysis method. First, a multidimensional parallelepiped uncertainty domain which takes into account the influence of parameter correlations is constructed. Second, the corresponding joint basic probability assignments are established for each focal element in the uncertainty domain. Finally, the reliability interval composed of the belief and plausibility measures are computed. Several numerical examples are investigated to demonstrate the effectiveness of the proposed model and the corresponding reliability analysis method.