Evidential reasoning rule for interval-valued belief structures combination

Article type: Research Article

Authors: Zhang, Xing-Xian^{a; c} | Wang, Ying-Ming^{a; b; *} | Chen, Sheng-Qun^d | Chu, Jun-Feng^a

Affiliations: [a] Decision Sciences Institute, Fuzhou University, Fuzhou, Fujian, P.R. China | [b] Key Laboratory of Spatial Data Mining & Information Sharing of Ministry of Education, Fuzhou University, Fuzhou, Fujian, P.R. China | [c] School of Architecture and Engineering, Tongling University, Tongling, Anhui, P.R. China | [d] School of Electronic Information Science, Fujian Jiangxia University, Fuzhou, Fujian, P.R. China

Correspondence: [*] Corresponding author. Ying-Ming Wang, Decision Sciences Institute, Fuzhou University, Fuzhou 350116, P.R. China. Tel.: +86 0591 22866677; E-mail: [email protected].

Abstract: Dempster-Shafer theory (DST) of evidence has wide application prospect in the fields of information aggregation and decision analysis. To solve the issues of interval evidence combination and normalization, we have reinvestigated the methods provided for interval evidence combination within the frameworks of DST and evidential reasoning (ER) approach, respectively, and pointed out the shortcomings of existing methods. A more general interval evidence combination approach based on the ER rule is constructed. Numerical examples are provided to indicate that the proposed method not only suitable to the conflict-free interval evidence combination, but also to the conflicting interval evidence combination, and interval evidence specificity can be kept intact in the interval evidence combination process. Moreover, the interval evidence combination methods based on DST or ER are special cases of the proposed method in some cases.

Keywords: Dempster-Shafer theory of evidence, interval evidence, ER approach, ER rule, combination

DOI: 10.3233/JIFS-182529

Journal: Journal of Intelligent & Fuzzy Systems, vol. 37, no. 2, pp. 2231-2242, 2019