Green Fog Computing Resource Allocation Using Joint Benders Decomposition, Dinkelbach Algorithm, and Modified Distributed Inner Convex Approximation
Ye Yu, Xiangyuan Bu, Kai Ming Yang, Zhu Han · 2018
Fog computing is a promising approach to alleviate the computation burden in traditional mobile networks to meet the increasing application demands. Such a complicated system is typically challenging and requires distributed solutions. In this paper, we investigate the resource allocation problem in fog computing to maximize the utility function from the energy efficiency perspective. The formulated problem is a mix integer nonlinear programming problem, which is NP-hard. We adopt a modified distributed inner convex approximation (NOVA) to approximate the problem first. Then, the Benders decomposition algorithm is applied to deal with integer variables. In the subproblem, we use the Dinkelbach algorithm to transform the fractional programming into an equivalent parametric subtractive form. Furthermore, the subproblem is decomposed distributedly, which enables users to update without information exchange. The simulation results indicate the effectiveness of the proposed algorithm.