对复杂系统的维修过程,需要构建一种有效的测试序列作为测试诊断策略,以获得高的故障隔离精度,并使包括测试费用、测试时间、测试所需的人力物力等在内的测试代价较低。针对系统维修中的顺序诊断策略问题,给出了问题的一般描述。对测试排序问题的最优算法,进行了算法过程分析并说明其在实现中存在的问题。在此基础上研究一种次优顺序诊断算法,即基于信息试探(IG)的诊断树产生算法,给出了根据系统相关矩阵进行算法设计的步骤,并将这种算法与最优算法进行了理论和实例仿真比较。结果表明基于信息试探的诊断树产生算法简单、实时性好,可用于复杂系统的在线诊断测试和维修诊断过程。

In maintenance procedures of complex systems, it is required to build an effective test sequence as the test/diagnosis strategy for obtaining higher fault isolation precision and less test costs on test expense, test time, manpower and material resources required.Directed towards the sequencing diagnosis problem in the system maintenance, the generic description of the system is presented.For the optimum algorithms of the test sequencing problem, the procedure of the algorithms is analyzed, and the problems existed in the applications of the algorithms are discussed.On which basis, a suboptimal sequential diagnosis algorithm, a diagnosis tree generation method based on IG, is presented.The steps for designing the algorithm according to the dependency matrix of the system are given.The method is compared with the optimal algorithm by theoretical analysis and simulation.The results demonstrated that the algorithm studied here is simple and has good real-time performance, which can be used in the on-line and off-line test and maintenance procedure for the complex systems.