Deep Reinforcement Learning Based Power control for Wireless Multicast Systems

TL;DR

This paper applies deep reinforcement learning with neural network function approximation to optimize power control in wireless multicast systems, effectively handling large state spaces and adapting to changing system conditions.

Contribution

It introduces a deep RL approach for power control in multicast systems, enabling near-optimal solutions in large, complex environments with dynamic adaptation.

Findings

Deep RL can learn effective power control policies for large multicast systems.

The method ensures average power constraints via learned Lagrange multipliers.

The approach adapts to time-varying system statistics with minor modifications.

Abstract

We consider a multicast scheme recently proposed for a wireless downlink in [1]. It was shown earlier that power control can significantly improve its performance. However for this system, obtaining optimal power control is intractable because of a very large state space. Therefore in this paper we use deep reinforcement learning where we use function approximation of the Q-function via a deep neural network. We show that optimal power control can be learnt for reasonably large systems via this approach. The average power constraint is ensured via a Lagrange multiplier, which is also learnt. Finally, we demonstrate that a slight modification of the learning algorithm allows the optimal control to track the time varying system statistics.