A Multi-objective Task Offloading Algorithm for Heterogeneous Cloud-Edge Collaboration in IoT

Cloud-edge collaboration has emerged as a promising architecture for executing Internet of Things (IoT) tasks with different requirements. Currently, the cloud-edge collaboration architecture assumes the cloud layer is not directly visible to terminals. The quality of service (QoS) evaluation metrics for cloud-edge task offloading studies mainly focus on latency, energy consumption, and cost. In this paper, a cloud-edge collaboration architecture including an offloading agent is presented, in which cloud-edge heterogeneous resources are treated as a logical resource pool. The agent assigns appropriate cloud or edge nodes to the tasks based on their characteristics. The task offloading problem is modeled as a multi-objective optimization problem with respect to energy consumption, computation time, and throughput rate. A genetic Algorithm based on the elite retention strategy is used to solve the problem. The results show that the algorithm improves the overall optimization performance compared to the baseline algorithm.