Digital twin-based optimization of AGV scheduling under energy constraints in plant factories

With the advancement of modern agriculture, plant factories have gained significant attention as an efficient mode of fruit and vegetable cultivation. In medium to large-scale plant factories, Automated Guided Vehicles (AGVs) are crucial for the transportation of produce. However, existing AGV scheduling systems often struggle with inefficiencies due to inaccurate energy consumption predictions and suboptimal scheduling algorithms. This paper addresses these challenges by proposing a digital twin-based approach to optimize AGV scheduling, focusing on improving the accuracy of AGV energy prediction models and enhancing scheduling efficiency. A digital twin framework for AGV scheduling in plant factories is developed, incorporating advanced single and multi-AGV energy consumption models. An improved Non-dominated Sorting Genetic Algorithm II (NSGA-II) is employed to design an energy-constrained AGV scheduling algorithm. Experimental results demonstrate a significant improvement in energy prediction accuracy and show that the proposed scheduling method reduces the maximum completion time of the AGV system by 8.025 % and decreases energy consumption by 7.62 % compared to traditional methods. This research provides robust technical support for cleaner production and sustainable development in plant factories.