A scheduling optimization method for AGV in automotive island assembly workshop using hybrid genetic algorithm

With the rising demand for customized automotive products, automotive island assembly workshops relying on Automated Guided Vehicles (AGVs) face significant challenges in scheduling efficiency. Existing methods often fail to adequately address dynamic production constraints, resulting in low resource utilization and extended cycles. To address these issues, this paper establishes a dual-resource scheduling model for AGVs and assembly islands. The model aims to minimize the total makespan by explicitly incorporating key operational characteristics and production constraints. We propose a Hybrid Genetic Algorithm (HGA) that integrates the A* algorithm with a simulated annealing mechanism. The method employs a task-module-based three-level real-number encoding structure to enhance scheme representation and search efficiency. Additionally, an improved crossover operator and simulated annealing are introduced to prevent premature convergence to local optima. During decoding, the A* algorithm combined with priority rules plans collision-free paths for AGVs. Simulation results demonstrate that the proposed HGA converges faster, effectively escapes local optima, and yields superior solutions, offering significant practical value for improving production efficiency in island assembly workshops.