In U-shaped automated container terminals (U-shaped ACTs), automated guided vehicles (AGVs) need to frequently interact with yard cranes (YCs), and separate scheduling of the two devices will affect terminal efficiency. Therefore, this study explores the coordinated scheduling problem between the two devices. To solve this problem, a high-precision simulation model of the U-shaped ACTs is established, which incorporates real operational logic. Second, an Improved Non-dominated Sorting Genetic Algorithm II based on Proximal Policy Optimization (INSGAII-PPO) is proposed. The algorithm uses PPO to realize dynamic genetic operator selection and makes related improvements, which improve the multi-objective optimization ability of NSGAII, and solve the collaborative scheduling problem by combining simulation. Finally, a hybrid weighted Technique for Order Preference by Similarity to Ideal Solution with preferences is proposed to select the final solution. The experimental results show that the scheme obtained by INSGAII-PPO exhibits better convergence and diversity, and offers significant advantages compared with the comparison algorithms. Moreover, the energy consumption and waiting time of the final solution selected by the proposed method are reduced by 3.42% and 4.87% on average. The proposed method has the capability of providing a theoretical reference for the AGVs and YCs collaborative scheduling of U-shaped ACTs.

Coordinating automated guided vehicles (AGVs) on shared roads is essential to efficient container-terminal transport. This study aims to improve routing and parameter adaptation in a U-shaped automated container terminal. We develop an improved ant colony optimization (IACO) algorithm combining direction-guided search, time-window conflict checks, and waiting or local replanning. A retrieval-augmented large language model recommends bounded parameter updates from domain literature and simulation experience, while IACO constructs and evaluates routes. In simulations on a 227-node network, IACO reduced mean makespan by 25.5–56.9% relative to conventional ant colony optimization, particle swarm optimization, and Rf-Ant, with shorter routes and fewer conflict-avoidance stops. A fleet-size sensitivity analysis examined 200–1200 transport tasks and 24–40 AGVs. At 100 tasks and 24 AGVs, ten tuning rounds across three scenarios selected parameters evaluated in 50 held-out scenarios; the 95% confidence interval for composite-score improvement was above zero. Retrieval yielded additional gains in most second-round task–fleet settings. Among four LLM backends, qwen3-max achieved the highest average improvement (3.60%) and beneficial-adjustment rate (73.6%). The approach supports conflict-aware routing and workload-sensitive fleet and parameter decisions in automated terminals.