An Efficient Fusion Algorithm on Conflicting Evidence

DSmT would resolve the issue of evidence portfolios when the high conflict of evidences appeared. But the computation oversized more easily because more focus elements were additional in the DSmT rules. And the fusion result was worse than that of DST when the low conflict situation was occurred. In order to efficiently combine high conflicting evidences, an improved fusion algorithm on conflict evidence is proposed. The method uses conflict mass as a basis for judgment in conflict situations. DST rules are adopted when conflict mass is lower. And DSmT fusion algorithm was adopted while opposition. During the switch between DST and DSmT, a mutual belief degree and support degree of the evidence is calculated firstly. Then the weight is obtained. Finally the conflict mass redistributed to every focus element weightily. Compared with the exist methods, this new algorithm has the advantage of better performance and more rational.