US 11,947,416 B2
Fault diagnosis in complex systems
Wei Xian, Xi'an (CN); Gang Lyu, Shanghai (CN); Dong Ping Song, Xi'an (CN); Geng Hu, Xi'an (CN); Yao Dong Zhang, Shanghai (CN); and Ke Qiang Chen, Beijing (CN)
Assigned to International Business Machines Corporation, Armonk, NY (US)
Filed by INTERNATIONAL BUSINESS MACHINES CORPORATION, Armonk, NY (US)
Filed on Dec. 14, 2021, as Appl. No. 17/644,227.
Prior Publication US 2023/0185653 A1, Jun. 15, 2023
Int. Cl. G06F 11/36 (2006.01); G06F 11/07 (2006.01); G06Q 10/0633 (2023.01)
CPC G06F 11/079 (2013.01) [G06F 11/0751 (2013.01); G06F 11/0778 (2013.01); G06F 11/3612 (2013.01); G06F 11/3616 (2013.01); G06F 11/3692 (2013.01); G06Q 10/0633 (2013.01)] 13 Claims
OG exemplary drawing
 
1. A method for identifying a weakness of a workflow in a complex system, comprising:
collecting runtime data about the complex system, the complex system comprising a plurality of subcomponents;
identifying an abnormal operation in the complex system;
constructing a multi-dimensional cause-and-effect relation matrix among the plurality of subcomponents, including assigning weights to each interaction of one of the plurality of subcomponents with respect to other subcomponents of the plurality of subcomponents with respect to each dimension of the multi-dimensional cause-and-effect matrix;
calculating a Most Likely Related Path (MLRP) score for the abnormal operation using the assigned weights;
filtering one or more related operations using the calculated MLRP score;
presenting the filtered one or more operations to a user;
determining if the MLRP score correctly inferred the real trouble area for the abnormal operation; and
in response to the determining, adjusting one or more of the assigned weights for one or more of the plurality of dimensions.