Abstract: Aiming at the task scheduling problem in automated warehouse with completion time constraint, a bin reassignment strategy is proposed to allocate goods to reasonable bins and generate task sequence constraints. A mathematical scheduling model is established with the objective of minimizing the total energy consumption of the stacker crane, and a corresponding penalty function is introduced. An improved sea-horse optimizer (I-SHO) algorithm is employed as the global optimization algorithm to solve this problem. Based on the sea-horse optimizer (SHO), the chaotic Tent mapping and opposition-based learning strategy are integrated to improve the quality of initial solutions while ensuring the ergodic uniformity. An adaptive t-distribution mutation strategy is introduced to increase population quality and prevent convergence to local optima in the algorithm. Besides, a correction mechanism is set up to ensure that solutions meet task sequence constraints. A hybrid population optimization strategy is introduced to further improve population quality and search accuracy. Finally, through the verification of examples, the sea-horse optimizer, genetic algorithm (GA), and particle swarm optimization (PSO) are used as comparison algorithms, and the feasibility and effectiveness of the I-SHO algorithm in solving the energy optimization scheduling problem of automated warehouse are verified.

Key words: task scheduling optimization, energy consumption optimization, correction mechanism, improved sea-horse optimizer algorithm

摘要： 针对带有完工时间约束的自动化立体仓库任务调度问题，提出一种货位再分配策略，对货物进行合理的货位分配，并产生任务先后约束，建立以堆垛机总能耗最低为优化目标的数学调度模型，并引入相应的惩罚函数，采用一种改进的海马优化算法（improved sea-horse optimizer，I-SHO）作为全局优化算法并进行求解。在原始海马优化算法（sea-horse optimizer，SHO）的基础上，融合混沌映射与对立学习策略，提高了初始解的质量。引入自适应t分布变异策略，避免陷入局部最优，并且设置修正机制，使解满足任务先后约束。引入混合种群寻优策略，进一步优化算法的搜索能力。最后通过实验进行验证，将海马优化算法、遗传算法（genetic algorithm，GA）和粒子群算法（particle swarm optimization，PSO）作为对比算法，验证了I-SHO在求解自动化立体仓库能耗优化调度问题上的有效性。

关键词: 任务调度优化, 能耗优化, 修正机制, 改进海马优化算法