计算机科学
服务器
移动边缘计算
边缘计算
分布式计算
调度(生产过程)
数据收集
最优化问题
任务(项目管理)
服务(商务)
GSM演进的增强数据速率
计算机网络
数学优化
人工智能
算法
经济
经济
管理
统计
数学
作者
Yanming Fu,Xian Zhang,Xiaoqiong Qin,Qingchang Meng,Bocheng Huang
标识
DOI:10.1016/j.comnet.2022.109551
摘要
As the number and type of mobile crowd sensing (MCS) data collection increases, more and more computation and processing are required, resulting in higher service cost and service delay, posing a huge challenge to traditional MCS. Currently, edge computing is being introduced to MCS to collect data to reduce service cost and service latency. In the offline mode, there are two issues that need to be addressed with the edge computing-based MCS data collection. First, for large-scale, multi-player, and multiple types of task data, edge servers are limited in computational resources and need to address issues such as task offloading and service cache scheduling. Second, in traditional MCS data collection, workers usually carry in a single type of task data, but it has now been proposed that workers need to carry multiple types of task data. To address the above problems, this paper proposes a joint optimization strategy for edge computing based on multi-player cooperative game and greedy differential evolution algorithm (MCG-GDE) to improve the service rate of edge servers and minimize the service cost and service latency in the data collection. We build a mathematical optimization problem for edge computing based on MCS data collection. The formulation of the optimization problem proves to be a NP-hard problem, so this optimization strategy constructs a task propagation scheme for multi-player cooperative games (MCG), where tasks carried by workers are reassigned to effectively reduce problem complexity and produce sub-optimal solutions to the mathematical model. Then, on the basis of the suboptimal solution, the optimal solution of the problem is obtained by the greedy differential evolution algorithm (GDE). Simulation results demonstrate that MCG-GDE outperforms other baseline strategies.
科研通智能强力驱动
Strongly Powered by AbleSci AI