基于Q-learning的移动群智感知任务分配算法  

Q-learning based sensing task assignment algorithm for mobile crowd sensing system

作　　者：胡华[1,2] 张强 胡海洋[1,2] 陈洁[1,2] 李忠金 HU Hua, ZHANG Qiang, HU Haiyang, CHEN Jie, LI Zhongjin(1. College of Computer Science, Hangzhou Dianzi University, Hangzhou 310018, China;2. Key Laboratory of Complex System Modeling and Simulation, Ministry of Education, Hangzhou Dianzi University, Hangzhou 310018, Chin)

机构地区：[1]杭州电子科技大学计算机学院,浙江杭州310018 [2]杭州电子科技大学复杂系统建模与仿真教育部重点实验室,浙江杭州310018

出　　处：《计算机集成制造系统》2018年第7期1774-1783,共10页Computer Integrated Manufacturing Systems

基　　金：国家自然科学基金资助项目(61572162;61272188;61702144);浙江省重点研发计划资助项目(2018C01012);浙江省自然科学基金资助项目(LQ17F020003)~~

摘　　要：移动群智感知环境中的任务分配是工作流研究领域中一个新方向,为解决应用任务在移动智能用户间的合理调度与分配,本文将机器学习中的Q-learning方法引入到工作流任务分配问题中,提出一种针对多目标的强化贪婪迭代方法。该算法从宏观层面上通过强化学习的每一次探索进行学习优化,微观层面上通过贪心算法为每一次迭代选择局部最优解,增强了算法的性能。对比其他3种算法,所提算法不但能降低时间和能耗开销,而且收敛速度较快,能够提高感知效率,可作为移动群体感知的工作流调度问题走向智能化的一种尝试。Task assignment in mobile crowdsensing environments has become a new research topic in the researching area on workflow systems. To solve the task assignment between mobile intelligent users reasonably, Q-learning thought was introduced into workflow scheduling, and a greedy algorithm based on Q-learning was designed. Without the need of knowing the detailed information of mobile users information and any other estimated assumptions, this algorithm exchanged with users continuously and learned from the environments to improve the quality of tasks assignment among mobile users. Simulation results showed that the proposed method could improve the perceptual efficiency and speed up the convergence. It had provided an intelligent attempt for the task allocation system in mobile crowdsensing environments.

关 键 词：移动群智感知 Q-learning方法 任务分配 算法 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]