User Satisfaction Oriented Resource Allocation for Fog Computing: A Mixed-Task Paradigm
Xincheng Chen, Yuchen Zhou, Long Xiang Yang, Lu Lv · IEEE Transactions on Communications · 2020
In this paper, we tackle the joint computation and communication resource allocation problem for mixed-task user-concerned fog computing. To deal with diverse kinds of computing tasks, we propose a mixed-task paradigm to support the co-existence of binary offloading and partial offloading. Considering the impact of users' satisfaction on fog computing service, we employ the user-weighted energy efficiency (UWEE) as the objective of resource allocation and develop a user-concerned mechanism (UCM) to sketch users' social features. Then, under the constraints of users' satisfaction, the resource allocation problem for UWEE maximization is formulated as a mixed integer nonlinear programming problem (MINLP), which cannot be properly solved by the traditional relaxation algorithms due to the co-existence of binary offloading and partial offloading. After transforming the problem with the replacement-based method, an augmented Lagrange method (ALM)-based resource allocation scheme is proposed to iteratively solve this joint optimization problem, in which AMSGrad is employed to accelerate the convergence. Simulation results demonstrate the superior performance of the ALM-based resource allocation scheme in terms of UWEE.