Overcoming overestimation characteristic to classical interval analysis
College
Engineering and Computer Science
Department
Ocean and Mechanical Engineering
Document Type
Article
Publication/Event/Conference Title
AIAA Journal
Publication Status
Version of Record
Abstract
Interval analysis is a widespread tool of analyzing uncertainty in mechanical and aerospace systems. The method was developed by mathematicians since 1960 to get bounds on rounding error and measurement errors in mathematical evaluation of expressions to develop numerical methods with reliable, guaranteed methods. Two cases were considered for the parameterized analysis. One related to fully dependent intervals, mimicking the case of special, functionally graded materials; and other related to intervals describing material distribution being independent of the rest of the parameters, but fully dependent of each other. It turns out that, in the latter case, the derived bounds are broader than in the former case. Both cases of parameterized treatment are associated with higher efficacy than the classical version of interval analysis. It should be noted that some other methods have been suggested in the literature to overcome the phenomenon of overestimation.
First Page
2093
Last Page
2097
DOI
10.2514/1.J053152
Publication Date
1-1-2014
Recommended Citation
Elishakoff, Isaac and Thakkar, K., "Overcoming overestimation characteristic to classical interval analysis" (2014). Faculty Scholarship. 319.
https://digitalcommons.fau.edu/faculty_papers/319