Privacy-preserving data analytics for smart decision-making energy systems in sustainable smart community

Nowadays, smart community becomes as a key component of Internet of Things (IoT). The fast booming of it significantly enhances the application of smart cities. To achieve data analytics and decision making for energy systems in smart cities, we propose use fog computing to establish a smart community due to the reduction of bandwidth consumption and latency. However, privacy issues are emerging in this scheme because continuous attacks put sensitive information under great threats. Besides, the loss of sensitive information degrade the quality of data analytics and decision making. Motivated by this, we propose a differentially private smart community model with game theory-based personalized privacy protection (GPDP). Differential privacy is deployed in a personalized way while a logarithmic function is leveraged to map data sensitivity to privacy protection level. Then we use game theory to model the confrontation and further derive optimized trade-off, which is denoted by the Nash Equilibrium. In addition, we develop a modified reinforcement learning algorithm to achieve fast convergence. The advanced evaluation results show the advantages of the proposed model from the perspective of the optimized trade-off between personalized privacy protection and data utility, which improves data analytics and decision-making performances for energy management in smart cities.

Automated summary

This paper proposes a differentially private smart community model with game theory-based personalized privacy protection (GPDP) and a modified reinforcement learning algorithm. The model optimizes the trade-off between personalized privacy protection and data utility, improving data analytics and decision-making performances for energy management in smart cities.