Dynamic Resource Scheduling Optimization under Industrial Internet-Enabled Smart Manufacturing Based on Deep Learning

Abstract

This study proposes a dynamic resource scheduling optimization method that integrates deep learning prediction to address the evolution and cooperative scheduling challenges of multiple types of dynamic resources in industrial internet-enabled manufacturing environments. First, a multidimensional input structure combining plan, state, and key performance indicator features is designed, and a Convolutional Neural Networks–Long Short-Term Memory-Attention based resource state prediction model is developed to jointly predict the future availability of two representative dynamic resources: machine and manpower. Second, based on the prediction results, a multi-objective scheduling optimization model is established with goals such as minimizing task completion time and maximizing resource utilization, and an improved Non-dominated Sorting Genetic Algorithm II is employed to achieve collaborative scheduling among multiple tasks and dynamic resources. Finally, simulation experiments are conducted in a representative intelligent manufacturing scenario to compare static, dynamic, and joint scheduling strategies. The results demonstrate that the proposed deep-learning-driven dynamic resource scheduling method effectively improves resource utilization, reducing completion time and task delays, thereby enhancing the responsiveness, and coordination of resource scheduling in industrial internet manufacturing systems. Furthermore, the obtained results provide a verifiable modeling and optimization framework for multi-resource collaborative scheduling in intelligent manufacturing.