Multiuser Computation Offloading and Resource Allocation for Cloud–Edge Heterogeneous Network
Qinglin Chen,Zhufang Kuang,Lian Zhao
DOI: https://doi.org/10.1109/jiot.2021.3100117
IF: 10.6
2022-03-01
IEEE Internet of Things Journal
Abstract:Cloud–edge heterogeneous network is an emerging technique built on edge infrastructure, which is based on the core of cloud computing technology and edge computing capabilities. The joint problem of computation offloading, cache decision, and resource allocation for cloud–edge heterogeneous network system is a challenging issue. In this article, we investigate the joint problem of computation offloading, cache decision, transmission power allocation, and CPU frequency allocation for cloud–edge heterogeneous network system with multiple independent tasks. The goal is to minimize the weighted sum cost of the execution delay and energy consumption while guaranteeing the transmission power and CPU frequency constraint of the tasks. The constraint of computing resource and cache capacity of each access point (AP) are considered as well. The formulated problem is a mixed-integer nonlinear optimization problem. In order to solve the formulated problem, we propose a two-level alternation method framework based on reinforcement learning (RL) and sequential quadratic programming (SQP). In the upper level, given the allocated transmission power and CPU frequency, the task offloading decision and cache decision problem is solved using the deep -network method. In the lower level, the optimal transmission power and CPU frequency allocation with the offloading decision and cache decision is obtained by using the SQP technique. Simulation results demonstrate that the proposed scheme achieves significant reduction on the sum cost compared to other baselines.
computer science, information systems,telecommunications,engineering, electrical & electronic