A Conflicting Evidence Fusion Method Based on a Greedy Strategy

To address the issue of the conventional D-S combination rule not effectively dealing with highly conflicting evidence fusion, a conflicting evidence fusion algorithm utilizing a greedy strategy is proposed. First, by adopting a greedy strategy, the effective evidence distance is defined based on Jousselme’s evidence distance, followed by the definition of strongly and weakly correlated evidence. The average value of the effective evidence distance is calculated, and based on this the evidence is classified. Finally, weights are calculated individually for the classified evidence. Weights are then reassigned to the two types of evidence, and the corrected evidence is fused to obtain the final fusion result. Numerical examples are provided to demonstrate the effectiveness of the method proposed in this paper in addressing the limitations of the conventional D-S combination rule for fusing highly conflicting evidence.