Importance measure of basic interval variable in non-probabilistic model

For basic interval variable in non-probabilistic model, a new non-probabilistic importance measure is presented to measure the effect of the basic interval variable on the non-probabilistic reliability index of system. The non-probabilistic importance measure can directly provide the global effect of the basic interval variables on the structure safety, thus it is able to guide engineering design. The main properties of the non-probabilistic importance measure are discussed, then Monte Carlo method and one-dimensional optimization algorithm are employed to estimate the non-probabilistic importance measure. The concept and the implementation of the presented method are described step by step, and several examples are used to demonstrate the efficiency and the precision of the presented method.