Research On Optimal Diagnosis Strategy Of Multi-Valued Attribute System

The optimal design of diagnostic strategy is an important research content in the system testability design,which can improve the reliability and stability of the system,and the test sequence is the key to the diagnostic strategy.At this stage,most of the research objects are binary attribute systems,and there are few studies on testing multi-valued attribute systems with multiple output values.The test sequence optimization problem of multi-valued attribute systems belongs to the category of combinatorial optimization problems.It can show excellent performance in such problems.This paper conducts in-depth research on the optimization of diagnostic strategies for multi-valued attribute systems.The specific work is as follows:Firstly,a quintuple mathematical model of the test sequence optimization problem is established by using the multi-signal flow graph,and the practical meaning of the multi-valued attribute system is studied and classified.According to the specific value range of element values in the correlation matrix,the difference between binary and multi-valued attribute systems is analyzed.According to the actual physical meaning of the elements in the multivalued matrix,the multivalued attribute system is divided into multivalued attribute system based on fuzzy group coding set,multivalued attribute system based on fault severity and multivalued attribute system based on test conditional probability.Secondly,an improved differential evolution algorithm(GCMDE)based on the GaussianCauchy mutation operator is proposed.By studying the characteristics of the Gaussian and Cauchy mutation operators,the two are fused with a differential strategy with different characteristics,simultaneously introducing parameters can attempt to update the iteration and the maximum un updated iteration to construct a complete algorithm system.According to the algorithm idea and the mutation operator ablation experiment,the rationality and optimization ability of the GCMDE algorithm can be preliminarily verified.Again,the performance of the GCMDE algorithm is studied in detail by unconstrained and constrained test functions.Fourteen standard unconstrained functions in CEC2005 and eighteen constrained functions in CEC2010 were used to test the algorithm,and compared with the differential evolution algorithm with excellent performance,the experimental results show that the GCMDE algorithm has good performance and global optimization ability.It can provide certain reference value for the improvement of differential evolution algorithm.Finally,according to the different diagnostic modes of the test sequence,the GCMDE algorithm is used to solve the diagnostic strategies of binary and multi-valued attribute systems.The scope of the test point presents two diagnostic modes and describes the steps for generating a fault tree.The diagnostic strategy of the binary attribute system is solved by the GCMDE algorithm,and compared with the optimal result,it contains fewer test points and lower test cost.When the GCMDE algorithm solves the diagnostic strategies of different types of multivalued attribute systems,the implementation of the algorithm is different,but the optimal test sequence can be obtained.For some instances,the algorithm generates a troubleshooting tree with fewer nodes and less spatial complexity.Experimental simulation shows that the GCMDE algorithm not only has significant advantages in solving the diagnostic strategy of binary attribute system,but also has certain feasibility and optimization performance for the diagnosis strategy problem of multi-valued attribute system.