A Network Capacity Optimization Strategy for Resource Constrained-UACNs

In Underwater Acoustic Communication Networks (UACNs), limited energy supply and complex, dynamic channels pose significant challenges for communication, and this has also become one of the research hotspots in the field of underwater acoustic communication in recent years. This paper presents a network capacity optimization strategy for power allocation in UACNs, that enables transmitters to learn optimal power levels based on the environment and the actions of other transmitters. The algorithm utilizes Reinforcement Learning (RL) to approximate the optimal action-value function and employs experience replay to improve learning efficiency. The simulation results demonstrate that the proposed power control strategy significantly outperforms traditional power control methods. Compared to random transmission power, the proposed algorithm triples the network capacity. In comparison to the continuous maximum transmission power strategy, the algorithm reduces energy consumption by 25%. Compared to a pre-defined strategies of transmission power, the algorithm better adapts to changes in the environment and interference. Achieving higher network capacity and lower energy consumption, the proposed algorithm is particularly effective in dynamic environments where the channel conditions and the actions of other transmitters change over time.