Integrated trucks assignment and scheduling problem with mixed service mode docks: A Q-learning based adaptive large neighborhood search algorithm

• Integrating dock mode, truck assignment, and scheduling improves operational efficiency • Unmanned warehouse operations benefit significantly from mixed service mode docks • Q-learning enhances operator selection efficiency in Adaptive Large Neighborhood Search • Optimal operators and selection mechanisms significantly improve algorithm performance Mixed service mode (MSM) docks enhance efficiency by flexibly handling both loading and unloading trucks in warehouses. However, existing research often predefines the number and location of MSM docks prior to planning truck assignment and sequencing. This predefined approach becomes less effective in high-demand systems and increases operational complexity, as warehouse operators must manually test various configurations. This paper addresses this challenge by proposing a new model that integrates dock mode decision, truck assignment, and scheduling, enabling more flexible dock mode arrangements. To solve the complex problem, we introduce a Q-learning-based adaptive large neighborhood search (Q-ALNS) algorithm, which adaptively adjusts dock modes through perturbation operators while simultaneously solving truck assignment and scheduling with destroy and repair operators. The Q-learning mechanism selects these operators based on their performance history and future gains, employing the epsilon-greedy strategy. Comprehensive experimental results and statistical analysis indicate that the Q-ALNS outperforms the benchmark in terms of optimality gap, with an average drop of 12.1%, while maintaining competitive computation efficiency. Compared to the predefined approach, our proposed adaptive strategy reduces tardiness by 22.5% and makespan by 7.6% on average, demonstrating its superiority in improving operational efficiency and supporting demand-driven assignment of MSM docks.