A mixed D-S combination rule based on the induced ordered weighted averaging operator and adaptive weighting

The D-S evidence theory is widely used in the fields of fusion fault diagnosis, multi-source information, safety assessment, etc. However, how to combine high-conflict evidence to obtain reasonable results remains a challenging problem. Based on this, this paper proposes a mixed D-S combination rule based on the Induced Ordered Weighted Averaging(IOWA) operator and adaptive weighting. Through numerical examples verification, in high-conflict situations, the accuracy of the proposed method compared to the Dempster method, Sun method, Abellan method, Murphy method, Tao method and Li method improved by 34.1%, 71.6%, 49.4%, 20.8%,14%, and 9.8% respectively. In low-conflict situations, the proposed method compared to the contrast methods improved by 5%, 67.8%, 13.8%, 5.7%, 5.3% and 5%, respectively. Next, the consumption time of the mixed D-S rule is in the same order of magnitude as D-S theory, Murphy method, Sun method and Tao method, while Abellan method and Li method is one order of magnitude higher than the mixed D-S rule. Finally, the proposed method are applied to the fault diagnosis of subway bogie bearings, In outer ring faults, the proposed method compared to the contrast methods improved by 7.1%, 63.9%, 43.1%, 9.1%, 6.2% and 3.2%, respectively. In inner ring faults, the proposed method compared to the contrast methods improved by 5.1%, 62.5%, 38.9%, 7.6%, 5.8% and 3%, respectively. The results show that the proposed method has certain advantages.