移动群智感知网络中用户参与在线激励机制研究

发布时间：2018-05-20 07:38

本文选题：移动群智感知网络 + 激励机制　；参考：《天津工业大学》2017年硕士论文

【摘要】：随着移动技术的发展,现有的移动终端集成了越来越多的传感器,每个移动终端都可以成为一个计算和信息资源交互设备,而大量的终端设备构建起的网络能够使用户快速、动态的获取和分享数据。这种感知方式的出现和普及,促进了移动群智感知网络的发展。移动群智感知网络"以用户为中心",充足的用户参与是移动群智感知网络的基础,因此如何吸引更多的用户参与到任务处理过程中去成为一个新的研究热点。本文在分析现有激励机制在用户选择的随机性、任务处理效率和预算控制方面的不足,提出两种用户参与激励机制策略,主要成果包括:提出了一种基于信誉模型的用户参与在线激励机制。该机制初始时为用户设置不同的信誉值,在任务分配过程中选择信誉度高的用户进行任务处理,并且通过因子调节用户的代价花费,最后对用户的信誉值进行更新之后用户可以进入到下一次的任务竞争中。通过实验验证,基于信誉模型的用户参与在线激励机制在提高用户参与率和任务完成率的同时,减少用户的花费,从而节省了总预算。提出一种基于用户阈值的用户参与在线激励机制。该机制首先将任务划分为三种类型,在用户收到任意类型的子任务时随机产生相应的参与阈值和一个对感知该子任务的预测花费,并上报给感知平台,然后感知平台根据这些信息选择合适用户进行任务处理,并计算出用户完成该子任务时应获得的报酬,最后用户利用特定方式决定是否参与该子任务的处理过程。通过实验验证,基于用户阈值激励策略在处理三种类型的子任务时,在提高任务完成的比例的同时,能够降低平台的预算花费。
[Abstract]:With the development of mobile technology, more and more sensors are integrated into the existing mobile terminals. Each mobile terminal can become a computing and information resource interactive device. Access and share data dynamically. The emergence and popularization of this kind of perception has promoted the development of mobile group intelligence perception network. The mobile group intelligence perception network is "user-centered", and sufficient user participation is the basis of mobile group intelligence perception network. Therefore, how to attract more users to participate in the process of task processing has become a new research hotspot. This paper analyzes the shortcomings of the existing incentive mechanism in user selection, task processing efficiency and budget control, and puts forward two strategies of user participation incentive mechanism. The main achievements are as follows: an online incentive mechanism for user participation based on reputation model is proposed. In the process of task assignment, the user with high reputation is selected to handle the task, and the cost of the user is adjusted by factor. Finally, the user's reputation value is updated and the user can enter the next task competition. The experimental results show that the online incentive mechanism of user participation based on reputation model not only improves the participation rate and task completion rate but also reduces the cost of users and thus saves the total budget. An online incentive mechanism for user participation based on user threshold is proposed. The mechanism divides the task into three types at first, when the user receives any kind of subtask, it randomly generates the corresponding participation threshold and a forecast cost of perceiving the subtask, and reports it to the perception platform. Then the platform selects the appropriate user to process the task according to the information, and calculates the reward that the user should get when he completes the subtask. Finally, the user decides whether to participate in the process of the subtask in a specific way. The experimental results show that the user threshold incentive strategy can reduce the budget cost of the platform while increasing the proportion of task completion while dealing with three types of subtasks.
【学位授予单位】：天津工业大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP212.9;TN929.5

【相似文献】