摘要
针对当前水库大坝灌区视频监控系统覆盖范围广、任务过多,导致系统时延过高的问题,提出一种基于模拟退火的自适应遗传禁忌搜索算法(Simulated-annealing-based Adaptive Genetic Taboo Search Algorithm,SAGTSA)的计算卸载策略。该策略通过改进染色体编码方式、自适应遗传算子和更新机制,并添加禁忌搜索的存储机制,使其跳出局部最优解的能力增强,从而系统能够根据当前时刻智能体设备的计算能力、任务的数据量大小以及任务数量等特点自适应地调整迁移至边缘节点或云中心的卸载策略。仿真实验表明,针对灌区视频监控系统场景下的业务需求,所提SAGTSA算法具有理想的时效性。
Aiming at the problems of wide coverage,too many tasks and high time delay of video surveillance system in reservoir dam irrigation area,a computational unloading strategy based on simulated annealing adaptive genetic tabu search algorithm(SAGTSA)is proposed. The strategy improves the chromosome encoding method,adaptive genetic operator and update mechanism,and adds the storage mechanism of tabu search to enhance its ability to jump out of the local optimal solution,so that the system can adaptively adjust the unloading strategy that migrates to the edge node or the cloud center according to the computing ability of the agent device at the current time,the amount of data and the number of tasks. Simulation results show that the proposed SAGTSA algorithm has ideal timeliness for the business requirements of video surveillance system in irrigation area.
作者
冯亚玲
张卓
张学武
徐晓龙
FENG Yaling;ZHANG Zhuo;ZHANG Xuewu;XU Xiaolong(College of Internet of Things Engineering,Hohai University,Changzhou 213022)
出处
《计算机与数字工程》
2022年第9期1917-1923,2047,共8页
Computer & Digital Engineering
基金
国家重点研发计划(编号:2018YFC0407101)
国家自然科学基金项目(编号:61671202)资助。
作者简介
冯亚玲,女,硕士研究生,研究方向:边缘计算;张卓,女,博士研究生,讲师,研究方向:智能视觉物联网理论与应用;张学武,男,博士研究生,教授,研究方向:传感网理论与应用、云计算与大数据;徐晓龙,男,硕士研究生,高级实验师,研究方向:智能水利技术与应用。
