Generative Adversarial LSTM Networks Learning for Resource Allocation in UAV-Served M2M Communications

This letter investigates the resource allocation problem for multiple Unmanned Aerial Vehicles (UAVs)-served Machine-to-Machine (M2M) communications. Our goal is to maximize sum-rate of UAVs-served M2M communications by jointly considering transmission power, mode, frequency spectrum, relay selection and trajectory UAVs. In order model uncertainty stochastic environments, we formulate be a Markov game, which generalization Decision Process (MDP) case agents. However, owning UAVs mobility poses difficulty perceiving environment, propose Long Short-Term Memory (LSTM) with Generative Adversarial Networks (GANs) framework better track forecast improving network reward. Numerical results demonstrate that proposed outperforms conventional LSTM Deep Q-Network (DQN) algorithms.