Answers are grounded in this paper and cite their sources.
Simulation study demonstrates dynamic path optimization in textile waste logistics networks, indicating reduced energy consumption and elevated recovery rates.
To improve the recycling efficiency of textile waste and reduce environmental pollution, this paper constructs a smart logistics system based on Deep Reinforcement Learning (DRL). Considering that large-scale intelligent logistics networks increasingly rely on distributed sensing, wireless communication, and reliable information transmission in electromagnetically complex environments, a Graph Neural Network (GNN) is employed to encode the state of a dynamic network composed of recycling points, vehicles, and sorting centers, capturing both node features and topological relationships. A Proximal Policy Optimization (PPO) algorithm is adopted to train multiple agents for collaborative path planning and task allocation. The model convergence is guided by a reward function that maximizes recycling benefits while minimizing energy consumption and idle distance. Experimental results demonstrate that the average recycling rate reaches 0.954, the task completion time is 65.5 minutes, and the energy consumption per unit of recycling is approximately 1.25 kWh/ton. The proposed method exhibits strong adaptability in dynamic environments and provides an efficient and feasible technical framework for intelligent textile waste recycling, while offering valuable insights for resource scheduling and cooperative optimization in communication-intensive smart systems.
No takes yet. Share an insight, caveat, or question.
Liu Z (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: