【摘要】 提出一种基于改进遗传算法(Improved Genetic Algorithm, IGA)的边缘计算任务卸载策略。针对边缘计算系统包含大量ECN(Edge Computing Node)和用户且任务具备结构化的特性，构建系统模型，并且详细分析用户延迟与能耗成本。将系统中的任务卸载决策转化成以成本最小化为优化目标的完全多项式非确定性(NP)问题。使用IGA求解NP问题，实现任务的高效卸载，其中使用整数编码、基于知识的交叉、种群分割的变异等操作提高算法的寻优能力。基于Python平台对所提策略进行实验论证，结果表明，种群数量设为35、迭代次数为350时，IGA的性能表现最好，且所提策略在用户设备总成本、ECN资源利用效率、执行时间等方面均优于其他对比方法。更多还原

【Abstract】 An edge computing task offloading strategy based on improved genetic algorithm(IGA) is proposed. Considering that the edge computing system contained a large number of edge computing nodes(ECN) and users, and the task was structured, the system model was constructed, and the user delay and energy consumption cost were analyzed in detail. The task offloading decision in the system was transformed into a complete polynomial nondeterministic(NP) problem with cost minimization as the optimization objective. IGA was used to solve NP problem to achieve efficient task offloading. Integer coding, knowledge-based crossover, and mutation of population segmentation were used to improve the optimization ability of IGA. The proposed strategy was demonstrated experimentally based on the Python platform. The results show that when the population size is set to 35 and the number of iterations is 350, the performance of IGA is the best, and the proposed strategy is superior to other comparison methods in terms of total user equipment cost, ECN resource utilization efficiency and execution time etc.更多还原