A Fish-swarm Algorithm Based Energy-saving Scheduling Strategy of Distributed Embedded Systems

摘要

Aiming at the real-time energy-saving scheduling problem in multiprocessor systems, a new scheduling strategy based on artificial fish-swarm algorithm is designed. Firstly, the height values of the task nodes are defined to determine the execution order of the task nodes. The task graph is split based on the critical path, the resulting paths are assigned based on the idea of load balancing, and then the task assignment scheme is obtained. Secondly, aiming at the multiprocessor system and its energy consumption model, an effective encoding strategy is proposed to determine the voltage regulation scheme. The deadline constraint is processed based on the feasibility rules to ensure the real-time performance of the system. On this basis, the corresponding problem model is established and solved by AFSA. The simulation results show that compared with the related scheduling strategies, the application of this scheduling strategy can achieve better load balancing, and then get better energy-saving effect.