A generalized weighted evidence fusion algorithm based on quantum modeling

In the field of evidence theory, the Dempster-Shafer framework is widely used for combining evidence. However, it often produces counterintuitive results when dealing with highly conflicting evidence. To mitigate this issue, the weighted evidence fusion method has gained popularity. However, with the generalization of the mass function from the perspective of quantum theory, the weighting of conflicting evidence under quantum effects remains an open issue. To address this issue, a generalized weighted evidence fusion algorithm based on quantum modeling is proposed in this study. First, when quantum interference effects are considered, Dempster's rule is generalized into a generalized evidence combination rule based on quantum modeling. Second, a generalized quantum evidence distance is proposed to measure the similarity of quantum evidence in the presence of phase-term interference. Third, a global and nonlinear quantum evidence weighting operator is designed to reduce the impact of conflicting quantum evidence. Then, based on this operator and the similarities between pieces of quantum evidence, the weighted quantum evidence is fused to yield the final conclusion using the proposed generalized combination rule. Finally, the rationality, reliability, and accuracy of the proposed method are effectively proven through experiments using a few numerical examples and applications.