Mobile crowdsensing has become a new way to perceive and collect information due to the widespread of smart devices. In this paper, we study the task assignment problem in mobile crowdsensing systems, which is aimed to reducing the average and largest makespan of all tasks. We consider scenarios where task requester needs the help of mobile users for task completion when they encounter directly or through AP cloud (i.e., several APs connected via wired/wireless links) in an opportunistic manner. We describe the mobile crowdsensing system and formulate the problems under study. We first derive the conditional expected encountering time between requester and different users by jointly considering the opportunities via direct encountering and indirect encountering via AP cloud. Then we propose an AP-assisted average makespan sensitive online task assignment (AP-AOTA) algorithm and an AP-assisted largest makespan sensitive online task assignment (AP-LOTA) algorithm. We present detailed design for both algorithms. We deduce the computational complexities of both algorithms to be O(mn²), where m represents the number of tasks and n represent the number of users. We conduct simulations on a real trace data set and a synthetic trace data set and the results show that our proposed algorithms significantly outperform existing work.

由于智能设备的普及，移动人群感知已经成为感知和收集信息的新方法。在本文中，我们研究了移动人群感应系统中的任务分配问题，目的是减少所有任务的平均和最大生成时间。我们考虑的场景中，任务请求者需要以机会方式直接或通过AP云（即，通过有线/无线链接连接的多个AP）遇到移动用户时，需要移动用户的帮助来完成任务。我们描述了移动人群感知系统，并提出了要研究的问题。我们首先通过共同考虑通过直接遇到和通过AP云间接遇到的机会，来得出请求者和不同用户之间的有条件预期遇到时间。然后我们提出了一种AP辅助的平均跨度敏感在线任务分配算法（AP-AOTA）和一种AP辅助的最大跨度敏感在线任务分配算法（AP-LOTA）。我们介绍两种算法的详细设计。我们推论两种算法的计算复杂度为O（m n ²），其中m表示任务数，n表示用户数。我们对真实的跟踪数据集和合成的跟踪数据集进行了仿真，结果表明，我们提出的算法明显优于现有工作。