IoT perception layer scheduling deadlock relieving optimization method

Abstract

When resources are limited and the perception tasks are time-constrained, if multiple perception tasks request the response at the same time, it is easy to cause scheduling deadlocks in the perception layer and affect its access speed and perception efficiency. This paper studies the mathematical feature expression method of task and resource allocation when deadlock occurs in the perception layer. According to the conditions of deadlock mutual exclusion, non preemption, request and hold, cycle waiting, a deadlock detection model of hierarchical scheduling system based on time constraints is adopted to establish the time constraint relationship of the deadlock detector and the genetic algorithm is applied to solve the problem of multi-tasks deadlock relieving strategy optimization in the perceptive layer. The research results and simulation experiments show that the deadlock relieving optimization method based on genetic algorithm can quickly relieve the deadlock, ensure the minimum cost of relieving, and significantly improve the efficiency of deadlock relieving.