Mobile Edge Caching Algorithm Based on User Location Prediction

‍ ‍The explosive growth of mobile traffic brought about by the development of the Internet of Things and 5G technology has greatly increased the burden of backhaul. Mobile edge caching technology can cache some files in the base station， which can avoid the repeated transmission of the content， and effectively relieve the burden of the backhaul. In order to improve the operator’s cache revenue， this paper proposes a multi-base station cache algorithm based on user position prediction. The algorithm uses the Long Short-Term Memory （LSTM） model to predict the user location. The model is trained by a large amount of real dataset， and the location information of each user is predicted. Then the user group within the service range of the base station at every moment is obtained. Based on the predicted information， combined with user preference information， the file request distribution is predicted. Under the constraint of limited base station cache capacity， the problem of maximizing the operator’s revenue is constructed. Combined with the file request distribution information， the global optimal cache strategy is obtained by dynamic programming algorithm. Simulation results show that the proposed algorithm can effectively reduce the backhaul cost and improve operator’s cache revenue.