Abstract: System-level diagnosis is an important approach to ensuring the reliability of multiprocessor computer systems. In order to improve diagnostic?ability and strengthen reliability, and motivated by the deficiency of the conditional diagnosability, peng et al. introduced a newly diagnosability, which is called g-good-neighbor conditional diagnosability. The g-good-neighbor conditional diagnosability is a new measure of system diagnosability, which is more suitable for massive multiprocessor computer systems. This paper takes exchanged crossed cube[(ECQ(s,t))] as the?object?of?study. By exploring the [Rg] vertex connectivity of [ECQ(s,t)], it determines, for the first time, the g-good-neighbor conditional diagnosability of [ECQ(s,t)] under the PMC model is [2g(s+2-g)-1] for [t≥s>g]. Furthermore, the correctness and validity of the conclusion are verified by simulation experiments. The research of [ECQ(s,t)] has important theoretical value and great practical significance for exploring the?reliable performance?of?[ECQ(s,t)] and promoting the application and popularization of [ECQ(s,t)].

Key words: exchanged crossed cube, [Rg] vertex connectivity, Preparata, Metze and Chien（PMC） model, g-good-neighbor conditional diagnosability

摘要： 系统级故障诊断是保障多处理器计算机系统运行可靠性的一种重要手段。为了提高系统的诊断能力，增强系统的可靠性，在条件诊断度的基础上Peng等人进一步提出了[g]正确邻结点条件诊断度，[g]正确邻结点条件诊断度是一种更加适用于大规模多处理器计算机系统的故障诊断方式。以新型互连网络拓扑结构研究的最新成果——交换交叉立方网络为研究对象，在得到交换交叉立方网络的[Rg]点连通度的基础上，首次证得交换交叉立方网络[(ECQ(s,t))]在PMC模型下的[g]正确邻结点条件诊断度为[2g(s+2－g)－1]，其中[t≥s>g]，进而通过模拟实验验证了结论的正确性和有效性。该研究对于理清交换交叉立方网络的可靠性能并有效推动交换交叉立方网络的应用和推广，有着非常重要的理论价值和现实意义。

关键词: 交换交叉立方网络, [Rg]点连通度, PMC模型, [g]正确邻结点条件诊断度