Multi-Path Routing Algorithm for WSN in Shipyard Based on MADRL

In shipyards, the transmission reliability and efficiency of wireless sensor networks are significantly affected by routing algorithms. Traditional routing algorithms are inefficient and prone to congestion. To address this challenge, this paper proposes a multi-path routing algorithm based on multi-agent deep reinforcement learning. The algorithm first designs the shortest multi-path routing strategy to enhance the flexibility of path selection; secondly, the stability of the algorithm is improved through the multi-agent deep reinforcement learning method. Through the evaluation of the network topology of an actual shipyard, the experimental results show that the proposed algorithm outperforms the existing benchmark methods in terms of stability and service quality, verifying its effectiveness and feasibility in practical applications.