系統偵錯(system-level diagnosis)是根據系統中各個處理器相互測試的結果， 推導出系統中錯誤處理器的程序。(t, k)偵錯是循序偵錯的概念化，當系統最多存 在t 個錯誤處理器時，在每一次進行偵錯時，至少須辨識出k 個錯誤處理器並且 以正常的處理器更換之。本篇論文假設每個處理器至少與一個正常處理器相鄰， 證明圖形在MM*模式下的(t, k)-條件式偵錯能力。 System-level diagnosis is used to identify faulty processors. Each processor in the system can test its neighboring processors and declare them fault-free or faulty. And (t, k)-diagnosis is an conceptualization of sequential diagnosis. In (t, k)-diagnosis, at least k faulty processors (or all faulty processors if fewer than k faulty processors remain) are identified and repaired in each iteration provided there are at most t faulty processors. In this paper, it assumes that each processor neighbor to at least one fault-free processor. We prove the conditional (t, k)-diagnosability of the r-regular graph under the MM* model.
