中文摘要：

诊断度是评估和判定多处理器计算机系统互连网络的自我故障诊断能力的重要指标。g好邻条件诊断度推广了经典的诊断度的概念,它要求每个非故障结点没有发生故障的邻点个数至少有g个。本文以PMC为模型,对排列图A_（n,k）的g好邻条件诊断度的上、下界进行讨论,最终得到了A_（n,k）在PMC模型下的g好邻条件诊断度为[（g＋1）k-g]（n-k）.

英文摘要：

It is an important indice for diagnosability to evaluate self diagnosis ability of the muhiprocessor system. G -good-neighbor conditional diagnosability popularized the concept of the traditional diagnosability, which restricts every fault-free node containing at least gfauh-free neighbors. In this paper, we show that theg -good-neighbor con- ditional diagnosability of An.k under the PMC model is [ （g ＋ 1 ）k - gl （n - k） by discussing the higher boundary and lower boundary of （ n, k） -arrangement graphs An,k respectively.