UAV deployment and caching scheme based on user preference prediction

In order to design an efficient edge caching policy considering spatial heterogeneity and temporal fluctuations of users' content requests, a proactive caching scheme was proposed with UAV's deployment location design based on user preference prediction. Firstly, each user's preference characteristics were predicted based on file similarity and user similarity, and the request time and user location were also predicted when a content request occurs. Thereafter, on the basis of the predicted geographical location, request time and user preference, each UAV's deployment location and the corresponding content placement were determined by virtue of clustering method based on SOM and AGNES. Simulation results show that the proposed scheme outperforms other three comparison schemes in terms of hit ratio and transmission delay. Furthermore, the results also reveal that content preference is correlated with different user features by different weights. Accordingly, different impact weights should be matched with different user features. © 2020, Editorial Board of Journal on Communications. All right reserved.