A Multi-task Real-Time Scheduling Algorithm Based on Dynamic Weight Compensation and Adaptive Task Jumping

Abstract

For a certain real-time task scheduling problems with characteristics of uncertain task sets, the executing time dynamic changing and the task's mutual coupling strength dynamic changing, a multi-tasking dynamic scheduling algorithm is proposed based on dynamic weight value compensation and adaptive task jumping, which solved the loads storming caused by the dynamic task scheduling as well as the too long system response time caused by the changes of the tasks' coupling coefficient. The scheduling diagram and function of the modules are given firstly. Then, the focuses are the implementation of the execution controller and the weight compensation method based on the research of task weight value, the weight compensation and weight classification. The reliability and efficiency of the algorithm is verified by comparing tests, and has been applied to the combining heavy-duty train's braking condition monitoring and early warning system.