Multiple Mix-Zones Deployment for Continuous Location Privacy Protection
Zhikai Xu, Hongli Zhang, Xiangzhan Yu · 2016
Privacy protection is critical for location-based services (LBS). A frequently proposed solution to protect location privacy suggests that the LBS users need to exchange their pseudonyms over time in protected regions called mix-zones. However, how to deploy mix-zones remains an open question and has not been well addressed in previous works. Inspired by the notion that "where you drive more is where you more likely need to change pseudonyms", we then treat the problem of optimal multiple mix-zones as a transportation problem. We then characterize the properties and constraints of the optimization problem, and build a mixed-integer programming model with the objective of minimizing the amount of time the users' privacy level is lower than the privacy requirement. The placement optimization problem is NP-hard. Therefore, we propose a heuristic algorithm to strategically select the mix-zone locations. Finally, we demonstrate the effectiveness of our solutions through extensive simulations on real-world data traces.