A Cache-Enabled Device-to-Device Approach Based on Deep Learning

In this paper, we present a deep learning-based Device-to-Device (D2D) approach that utilizes Gated Recurrent Unit (GRU) model is optimized through Bayesian optimization for hyperparameter tuning. The proposed approach, DLCE-D2D (Deep Learning Cache-Enabled device-to-device) system learning using GRU, to predict the popularity of content in D2D network and dynamically adjusts cache eviction policy improve hit ratio. was evaluated real-world data compared against traditional policies such as Least Recently Used (LRU) First Out (FIFO). results show outperforms terms Also, it demonstrated robust changes access patterns can adapt dynamic network. Additionally, shown GRU achieve similar or better than other methods. use tuning offers promising solution improving ratio networks, thereby performance reducing cost networks.