Reinforcement learning based resource allocation in cache-enabled small cell networks with mobile users

Abstract

In this paper, the resource allocation problem for cloud-based cache-enabled small cell networks (SCNs) is studied. In the cloud-based cache-enabled SCN, the contents that the users request are stored both at the cloud pool and at the cache storage of each small base station (SBS). In our model, the cloud pool can predict the users' mobility patterns and determine the resource allocation scheme in a period of time. The problem is formulated as a game problem which jointly considers the network throughput and the power consumption of the SBSs. To solve this problem, we propose a machine learning based resource allocation method. First, we use a neural network framework of long short-term memory to predict the users' mobility patterns and further to determine the associated users based on the users' mobility patterns. Then we propose a reinforcement learning based resource allocation algorithm to maximize the network throughput. Simulation results show that the proposed algorithm achieves up to 58.2% and 26.1% gains, respectively, in terms of network throughput compared to random and the nearest algorithms.