Unsupervised Learning Neural-Network Method for Resource Allocation in Multi-cell Cellular Networks

Abstract

In order to meet requirements of high speed, low energy consumption and low latency of resource allocation in multi-cell cellular networks, an unsupervised learning based resource allocation method is proposed. The unsupervised learning based resource allocation method contains an unsupervised learning based power control neural network (NNPC) and an unsupervised learning based channel allocation neural network (NNCA). The NNPC is constructed to output optimized channel power to maximize the expectation of linear weighted sum of transmission rate and energy efficiency, and the NNCA is constructed to combine the NNPC to further optimize the expectation of linear weighted sum of transmission rate and energy efficiency. The simulation results show that the proposed method can obtain better transmission rate and energy efficiency than other algorithms under the premise of ensuring low delay.