Linkage Mechanism of Employee Performance Optimization and Functional Textile Quality Control Based on Improved DDQN

To address the challenge of jointly optimizing employee performance and functional textile quality in intelligent manufacturing environments, this study proposes a multi-objective collaborative optimization framework based on an improved Double Deep Q-Network (DDQN). A multidimensional state-space model is constructed by integrating employee behavioral indicators, production process parameters, and real-time quality inspection data, enabling coupled representation of human–machine interactions and manufacturing performance. The proposed method incorporates a dynamic reward function, prioritized experience replay, and multi-step temporal difference prediction to improve policy learning efficiency and long-term decision stability. Considering that modern textile manufacturing increasingly relies on heterogeneous sensing devices and distributed industrial information acquisition, the proposed state modeling and decision-making framework provides a scalable architecture for integrating multimodal production data and supporting intelligent monitoring systems. Experimental evaluations demonstrate that the improved DDQN reduces the average defect rate to 2.6% with a standard deviation of 0.4 while achieving a comprehensive performance score of 89.4, significantly outperforming conventional DDQN and static scheduling strategies. The framework effectively balances production efficiency, operational consistency, and quality stability, offering a practical solution for intelligent textile manufacturing and providing methodological insights for future interdisciplinary research on advanced sensing, industrial information fusion, and smart manufacturing infrastructures.