Affiliations: [a] College of Mathematics and Statistics, Nanyang Institute of Technology, Nanyang, China
| [b] Spatial Epidemiol Lab SpELL, University Libre Bruxelles, Brussels, Belgium
Correspondence:
[*]
Corresponding author. Yunxu Tong, College of Mathematics and Statistics, Nanyang Institute of Technology, Nanyang 473004, China. E-mail: lgh_tong@163.com.
Abstract: The current method does not take full account of multiple hybrid tasks in fuzzy control system and the problems of the balance between the requirements of the system reliability and the maximum completion time of the scheduling and the high cost of resource occupancy. To address these problems, a fault-tolerant scheduling algorithm of multiple hybrid tasks based on supporting multilevel criticality is proposed in this paper. The models of fuzzy control system and multiple hybrid tasks are built respectively. Multiple hybrid tasks in the model are divided into periodic task and non-periodic task, task with fault-tolerant requirement and with no fault tolerance requirement. According to the priority of each task and the relationship of the response time and time limit of each task, whether to start its supplementation task and the fault tolerance priority allocation is determined. The worst response time of each task in the model is calculated and fault-tolerant scheduling for multiple hybrid tasks is realized. Experimental results show that the proposed algorithm can further reduce the maximum completion time of task scheduling on the basis of satisfying the reliability requirements of the fuzzy control system. The cost of computer resource occupancy and the overhead of communication resources have been greatly reduced.
