To solve the problem of limited computing ability in mobile devices, edge computing is adopted as a feasible solution which provides services for IoT devices in different geographical locations. However, due to the service uncertainties, including the network congestion and the performance degradation of edge nodes, novel offloading strategies must be developed to accommodate the uncertain situations. In view of this challenge, software-defined network (SDN) is integrated with edge computing to make service offloading more flexible. Technically, an optimal IoT service offloading (OSO) method with uncertainty is proposed. In OSO, the completion time and load balance variance are two optimization goals for developing offloading strategies, and then the non-dominated sorting genetic algorithm-II (NSGA-II) is fully investigated to improve the performance in completion time and load balance variance. Moreover, the optimal strategy is selected by using Simple Additive Weighting (SAW) and Multiple Criteria Decision Making (MCDW). Finally, the experimental evaluation is conducted by comparing OSO with other methods to verify the superiority of it.

为了解决移动设备计算能力有限的问题，边缘计算被作为一种可行的解决方案，为不同地理位置的物联网设备提供服务。然而，由于服务的不确定性，包括网络拥塞和边缘节点的性能下降，必须开发新的卸载策略来适应不确定的情况。针对这一挑战，软件定义网络（SDN）与边缘计算相结合，使业务分流更加灵活。从技术上讲，提出了一种具有不确定性的最优物联网服务卸载（OSO）方法。在 OSO 中，完成时间和负载均衡方差是开发卸载策略的两个优化目标，然后充分研究了非支配排序遗传算法-II（NSGA-II）以提高完成时间和负载平衡方差的性能。此外，通过使用简单加法加权（SAW）和多准则决策（MCDW）来选择最优策略。最后，通过OSO与其他方法的比较进行实验评估，以验证其优越性。