Fault Diagnosis Algorithm For Interconnection Network System

With the advancement of science and technology,many subject areas have unlimited desire for high computing power,such as deep learning,simulation of atmospheric and ocean currents,computational chemistry,aerodynamics,etc.,all relying on multi-processor computers.A key aspect of maintaining the availability of a multiprocessor computer system with a complex interconnected network structure is system level fault diagnosis of the interconnected network.In the study of system-level fault diagnosis,scholars have unearthed many diagnostic models,but high-quality diagnostic algorithms are not common.In this regard,this paper proposes two novel fault diagnosis algorithms.Hypercube network is a specific network topology with a wide range of applications.To solve the problem of high complexity of existing fault diagnosis algorithm in hypercube network,the concept of fault fan was introduced,the parallel depth-first search strategy algorithm was used to find the fault fan in hypercube networks.And the fault node of the network was determined in order to replace or repair it,which provides an important new way for the reliability of the network.This paper provided a significant new way for the reliability of the network.The time complexity of this algorithm is not great than that of the existing algorithm with complexity greater than the square.Hypercube network is only a special case of t-diagnosable system.In order to solve the more general t-diagnosable diagnostic problems of large-scale networks,we transfer the problem of fault diagnosis of a t-diagnosable system into the problem of the minimum cover of a so-called decision matrix A,which is equal to the problem of zero-one integer program.Then,we design an improved evolution algorithm to solve the problem of zero-one integer program.Moreover,we obtain the fault set of the t-diagnosable system.