@article{sun-luo-tan-huy-zho-16-aa-reliab,
  author = {Jianping Sun and Yiping Luo and Zhaoping Tang and Yutao Hu and Xinjian Zhou},
  title = {An Improved Affine Algorithm for the Non-Probabilistic Reliability Index of Interval Model},
  journal = {International Journal of Control and Automation},
  issn = {2005-4297},
  volume = {9},
  number = {4},
  pages = {333-346},
  year = 2016,
  url = {https://www.earticle.net/Article/A272700},
  abstract = {Aimed at that standard interval arithmetic is easy to cause error explosion, as well as traditional affine arithmetic is easy to cause calculation error due to not considering the correlation among the first degree item, quadratic term, even the multiple-item and the nonlinear term expressed by the same noise, this paper proposed an improved algorithm which took affine algorithm operation as the main body. When there are those correlation items, the algorithm calculates the upper and lower bound of their total value by interval arithmetic, then uses the bound information and introduces the new noise affine form to instead of those terms. The method improved the conventional affine operation by combining interval arithmetic. When it applied to calculate the nonprobabilistic reliability index, the calculation results has higher accuracy than the interval arithmetic and the conventional affine algorithm, and it has strong practical significance.}
}