Deep reinforcement learning-based optimization of steel bridge deck paving process parameters
基于深度强化学习的钢桥面铺装施工工艺参数优化

摘要

钢桥面铺装施工工艺参数对铺装质量与耐久性影响重大。本研究聚焦于将深度强化学习技术引入钢桥面铺装施工工艺参数优化领域。通过构建精准的深度强化学习模型，对包括温度控制、碾压遍数、材料配比等关键施工工艺参数进行系统优化。详细分析了模型在不同工况下的学习与决策过程，验证其能有效提升钢桥面铺装施工工艺参数的合理性，显著改善铺装质量，降低施工成本，提高施工效率。深度强化学习为钢桥面铺装施工工艺参数优化提供了创新性且高效的解决途径，具有重要的理论与实践意义。

Abstract

The construction process parameters of steel bridge deck paving significantly influence paving quality and durability. This study focuses on applying deep reinforcement learning technology to optimize these parameters. By developing a precise deep reinforcement learning model, we systematically optimized key construction parameters including temperature control, compaction passes, and material mix ratios. Through detailed analysis of the model's learning and decision-making processes under different working conditions, it was verified that this approach effectively enhances the rationality of construction parameters, significantly improves paving quality, reduces construction costs, and increases efficiency. Deep reinforcement learning provides an innovative and efficient solution for optimizing steel bridge deck paving process parameters, demonstrating significant theoretical and practical implications.