车联网边缘计算场景下基于改进型NSGA-Ⅱ算法的边缘服务器部署决策

Edge server deployment decision based on improved NSGA-Ⅱin the Internet of vehicles edge computing scenario

车联网环境下,边缘服务器的放置位置与部署数量直接影响到边缘计算的效率。由于在宏基站或基站上部署大型边缘服务器的成本较高,可以在微基站上部署一个小型边缘服务器作为补充,并通过优化大型边缘服务器的放置位置来降低成本。为了最小化边缘服务器的部署代价和服务延迟、最大化运营商的收入和服务器负载均衡度,把边缘服务器放置问题与车联网用户应用服务放置问题联合建模为一个多目标优化问题,并提出了基于改进型NSGA-Ⅱ算法的放置方案。实验结果表明,提出的边缘服务器放置方案能够降低约44%的边缘服务器部署成本,降低约14.2%的时延,提升24.2%的运营商收入,具有较好的应用价值。

In the context of the Internet of vehicles,the placement and deployment number of edge servers directly affect the efficiency of edge computing.Due to the high cost of deploying a large edge server on a macro base station and a base station,it can be complemented by deploying a small edge server on a micro base station,and the cost reduction needs to be optimized by optimizing the placement of large edge servers.In order to minimize the deployment cost and service delay of the edge server,and maximize the operator’s revenue and server load balance,the edge server placement problem combined with the vehicle networking user application service was modeled as a multi-objective optimization problem and a placement scheme based on improved NSGA-Ⅱalgorithm was proposed.The experimental results show that the proposed scheme can reduce the deployment cost of edge servers by about 44%,the latency by about 14.2%,and improve the revenue of operators by 24.2%,which has good application value.