An early warning method based on fuzzy evidential reasoning considering heterogeneous information

Abstract

Early warning is an effective way to eliminate or reduce the losses and negative impacts caused by hazards in advance, which has important practical significance in real world situations. It has drawn great attention from government and academy in recent years. Information related to hazards plays an important role in early warning process, there are different types of such information in real world situations. However, existing early warning studies employ just single information type to describe the hazards, it might not describe the hazards properly and comprehensively due to the complex and uncertain of real world situations. Therefore, it seems necessary and reasonable to consider different types of information related to hazards in the early warning process. To fill up the gap, this study proposes a novel early warning method that considers heterogeneous information from both quantitative and qualitative contexts including experts’ hesitancy. On such basis, fuzzy evidential reasoning approach is utilized due to its capacity not only to fuse different types of information, but also to keep as much information as possible during the fusion process. Finally, an illustrate example together with related discussions are presented to demonstrate the validity, feasibility and novelty of the proposed method. The proposed early warning method can provide more safety status information of early warning objectives, which can be used to not only forecast and warn different types of hazards, but also conduct the risk analysis and assessment. It has widely potential application values in real world situations.