Optimization of automated guided vehicle scheduling model based on hybrid genetic algorithm and digital twin

• The path length obtained by HGA is the shortest, 298m, with the shortest running time of 26.33s and an average time of 27.19s, which is 3.33s shorter than the bat algorithm and 2.86s shorter than GA. • The running time of the AGV dynamic scheduling method based on digital twins is 2614 seconds, with an empty vehicle rate of 38.3%. Compared with traditional static scheduling methods, the running time is reduced by 887 seconds, and the empty vehicle rate is reduced by 9.1%. • The IPSO algorithm converges rapidly in the early stage, stabilizes after 20 iterations, and can jump out of local optima in the later stage, with a convergence speed increase of about 40%, indicating that the Circle chaotic mapping strategy can effectively improve initialization quality. The scheduling of automated guided vehicles faces issues due to unpredictable data. To overcome this, research focuses on optimizing path planning and scheduling using hybrid genetic algorithms and digital twins. Firstly, a genetic algorithm improves the bat algorithm to avoid premature convergence to local optima, constructing a path planning model. Additionally, a scheduling optimization model based on digital twins is developed. This study used SolidWorks software to construct a 3D digital twin model of the warehouse environment, which synchronized the automated guided vehicles status in real-time through 10Hz sensor data and mapped it to the virtual model. Results show that the hybrid genetic algorithm achieves a shortest path of 31.88m and an average path of 32.31m, with a runtime of 26.33s (27.19s on average), outperforming the bat algorithm by 3.33s and the genetic algorithm by 2.86s. The dynamic scheduling method based on digital twins runs for 2614s with a 38.3% empty car rate. Furthermore, the proposed improved particle swarm optimization algorithm exhibits fast convergence, avoiding local optima. Overall, this model effectively enhances the loading and unloading efficiency in warehousing or port logistics, promoting industry development. To optimize AGV scheduling, this study proposes a path planning model based on HGA and a scheduling optimization model based on digital twins. The experiment proved that the shortest path of the proposed model is 31.88m, the average path is 32.31m, the running time is 2614s, and the empty car rate is 38.3%, which is lower than the static scheduling method.