基于移动社交网络的群智感知社群化任务分发

针对传统的“平台-用户”移动群智感知任务分发方式在任务执行鲁棒性以及数据回收方面存在的缺陷与不足，以群智感知参与用户的群体社会属性为基础，提出一种“平台-社群-用户”的社群化任务分发方法.基于参与用户时空移动特征分布的近似度计算对用户进行动态社群划分聚类，通过在社群中设置社群组织者与社群从属者角色，同时引入社交亲密度连接关系网络等模型，构建移动群智感知任务初次社群分发与二次用户选择的方法.基于WTD公开数据集上的实验结果表明，与传统的任务分发方式相比，所提方法可可有效提升任务完成率，增强执行过程的鲁棒性，同时可减少任务分发次数、降低通信负载.
Abstract: A "platform-community-user" socialization task allocation approach was proposed by leveraging participant users' group social characteristics in view of the shortcomings of traditional "platform-user" mobile crowd sensing model. Firstly, a dynamic community division clustering was achieved based on the similarity of users' spatiotemporal mobility distribution. And then, via the roles of the community organizer and subordinate, a novel mobile crowd sensing task mechanism, i.e., primary community distribution and secondary user selection, was proposed by incorporating social intimacy connection network. The experimental results on public WTD data set show that, compared with the traditional task allocation method, the proposed method can effectively improve the task completion rate, enhance the robustness of the execution process, reduce the number of execution iteration and the communication overhead.