返回
Optimizing task allocation with temporal-spatial privacy protection in mobile crowdsensing
DOI:10.1111/exsy.13717.png)
摘要
En 中文
Mobile Crowdsensing (MCS) is considered to be a key emerging example of a smart city, which combines the wisdom of dynamic people with mobile devices to provide distributed, ubiquitous services and applications. In MCS, each worker tends to complete as many tasks as possible within the limited idle time to obtain higher income, while completing a task may require the worker to move to the specific location of the task and perform continuous sensing. Thus the time and location information of each worker is necessary for an efficient task allocation mechanism. However, submitting the time and location information of the workers to the system raises several privacy concerns, making it significant to protect both the temporal and spatial privacy of workers in MCS. In this article, we propose the Task Allocation with Temporal-Spatial Privacy Protection (TASP) problem, aiming to maximize the total worker income to further improve the workers' motivation in executing tasks and the platform's utility, which is proved to be NP-hard. We adopt differential privacy technology to introduce Laplace noise into the location and time information of workers, after which we propose the Improved Genetic Algorithm (SPGA) and the Clone-Enhanced Genetic Algorithm (SPCGA), to solve the TASP problem. Experimental results on two real-world datasets verify the effectiveness of the proposed SPGA and SPCGA with the required personalized privacy protection.
Keyword:
differential privacy
mobile crowdsensing
privacy protection
task allocation
期刊
IF:
2.3
论文数:
2.6K
被引数:
3.8K
机构
引用论文
Dexterous robotic manipulation using deep reinforcement learning and knowledge transfer for complex sparse reward-based tasks使用深度强化学习和知识转移的灵巧机器人操纵复杂的稀疏基于奖励的任务
EXPERT SYSTEMS
IF2.3

