An Efficient Computation Offloading Management Scheme in the Densely Deployed Small Cell Networks With Mobile Edge Computing

Fengxian Guo, Heli Zhang, Hong Ji, Xi Li, Victor C. M. Leung

IEEE/ACM Transactions on Networking · 2018 · 291 citations · 38 references

Concepts

TL;DR

Mobile edge computing enables UEs to offload computation to nearby servers, and its integration into small cell networks is natural, yet comprehensive studies on efficient offloading in densely deployed SCNs are lacking. The study aims to minimize UE energy consumption by jointly optimizing offloading decisions, spectrum, power, and computation resource allocation in densely deployed small cell networks. We formulate the problem as a mixed‑integer nonlinear program and solve it with a hierarchical genetic‑algorithm and particle‑swarm‑optimization based sub‑optimal algorithm that determines offloading decisions and resource allocations. Simulation results show the algorithm converges and outperforms baseline methods.

Abstract

To tackle the contradiction between the computation intensive applications and the resource-hungry mobile user equipments (UEs), mobile edge computing (MEC) has been provisioned as a promising solution, which enables the UEs to offload the tasks to the MEC servers. Considering the characteristics of small cell networks (SCNs), integrating MEC into SCNs is natural. But in terms of the high interference, multi-access property, and limited resources of small cell base stations (SBSs), an efficient computation offloading scheme is essential. However, there still lack comprehensive studies on this problem in the densely deployed SCNs. In this paper, we study the energy-efficient computation offloading management scheme in the MEC system with SCNs. The aim of this paper is to minimize the energy consumption of all UEs via jointly optimizing computation offloading decision making, spectrum, power, and computation resource allocation. Specially, the UEs need not only to decide whether to offload but also to determine where to offload. First, we present the computation offloading model and formulate this problem as a mix integer non-linear programming problem, which is NP-hard. Taking advantages of genetic algorithm (GA) and particle swarm optimization (PSO), we design a suboptimal algorithm named as hierarchical GA and PSO-based computation algorithm to solve this problem. Finally, the convergence of this algorithm is studied by simulation, and the performance of the proposed algorithm is verified by comparing with the other baseline algorithms.

References

38