Genetic Algorithm based Edge Computing Scheduling Strategy
Chang Su, Yining Gang, Chengming Jin · 2021
With the rapid development of Internet and cloud computing, edge computing (EC) as a new technology has been constantly integrated into people's lives. Users can offload the tasks on the device to the nearby edge server for scheduling under EC, so as to get low latency and high efficiency service. Thus, how to schedule the tasks offloaded to the edge server has become an important problem. In order to solve the task scheduling problem under EC, we propose a task scheduling strategy for EC based on the improved non-dominated sorting genetic algorithm (Im-NSGA-II). The adaptive strategy of dynamic adjustment of crossover rate and mutation rate is used to improve the efficiency of population search, which is improved under the condition of maintaining the diversity of the population, and finally a group of optimal solution sets can be obtained. Compare with the traditional NSGA-II algorithm and other genetic algorithms, Im-NSGA-II has a great improvement in performance. The experimental results indicate that the proposed Im-NSGA-II algorithm has shorter task completion time compared with other genetic algorithms. In addition, it also has low latency, which indicates that the algorithm has better quality of service and better task scheduling effect.