Prediction of entering processes into the deadlock state

Andrii Nicheporuk (Khmelnytskyi National University)
Yuriy Klots (Khmelnytskyi National University)
Oksana Yashyna (Khmelnytskyi National University)
Sergiy Mostovyi (Khmelnytskyi National University)
Yuriy Nicheporuk (Khmelnytskyi National University)

Indonesian Journal of Electrical Engineering and Computer Science
Vol 14, No 3: June 2019

The work is devoted to the problem of processes’ deadlock in the multi-tasking system. On the basis of the study of the life cycle of the process, the boundary state of the process was distinguished, which will precede the deadlock state. Proposed the method for prediction of entering processes into the deadlock state, which consists of two algorithms: algorithm of detection of potential processes that may fall into the deadlock and algorithm of processes detection that falling into the state of deadlock. An evaluation of time complexity of proposed method is conducted. Unlike the known methods and algorithms, proposed method use fuzzy logic components to detect two or more processes that fall into the state of deadlock, and thus do not make the algorithm cumbersome, which allows it to be used in modern operating systems.