An interval construction belief rule base with interpretability for complex systems

Complex system modeling using only subjective judgment is insufficient to address the needs of more extensive and complex actual systems. The expert knowledge base in the belief rule base (BRB) take part in complex system modeling well, but there are still the following problems. On the one hand, BRB cannot deal with the overabundance of combination rules caused by the excessive number of complex system attributes. On the other hand, parameters may lose their interpretability after BRB modeling, reasoning, and optimization. Given the above two problems, an interval construction belief rule base with interpretability for complex systems (IBRB-i) model is proposed. The IBRB-i model uses interval addition to create belief tables and adds rule reliability to prevent combination rule explosion in traditional BRB. In addition, interpretable constraints are added to the IBRB-i model to ensure its interpretability. In the case study, the validity and precision of the IBRB-i are analyzed and verified by two examples of power fault detection and liquid launch vehicle's structure safety assessment.

Automated summary

Modeling complex systems using only subjective judgment is insufficient. The expert knowledge in the belief rule base (BRB) is useful, but it has problems dealing with excessive combination rules caused by complex system attributes and losing interpretability of parameters. To address these issues, an interval construction belief rule base with interpretability for complex systems (IBRB-i) model is proposed. The IBRB-i model uses interval addition to create belief tables, adds rule reliability to prevent combination rule explosion, and incorporates interpretable constraints. The validity and precision of the IBRB-i model are analyzed and verified through case studies on power fault detection and liquid launch vehicle's structure safety assessment.