自动化立体仓库中货位实时分配优化问题研究

Optimization of Real-Time Storage Location Assignment in Automated Storage and Retrieval System

本文研究了单元货格式自动化立体仓库中的货位实时分配问题。该问题可分为两方面:为入库分配空货位和为出库选择货位。在考虑了堆垛机的加减速对其运行速度的影响后,以堆垛机将要进行的所有操作的行程时间之和作为优化目标,通过调整堆垛机的后续操作的行程时间在总行程时间中的权重,依次构建了三个优化目标函数。然后利用结合了模拟退火方法的遗传算法对其进行了求解,并利用了多种技术提高算法的效率。最后利用仿真技术,通过改变仿真时系统中的货物类型总数、堆垛机的停留策略和装载能力,对该算法在不同情况下的效果进行了验证。仿真结果表明在不同情况下,该算法都可不同程度地减少堆垛机的平均行程时间。

This paper concerns storage location assignment in unit load automated storage/retrieval system. It can be divided into two aspects： assign the bin for the storage unit; choose the bin for retrieval command. With the consideration of the impact of crane＇ s acceleration and deceleration on crane＇ s ve- locity, the total travel time of the crane＇ s current operation and subsequent operations is used as the optimization object, then three optimization goal functions are developed by changing the weights of the crane＇s subsequent operations. The problem is solved by the genetic algorithm combined with the simulated annealing while the kinds of methods are used to improve the algorithm＇s efficiency. The algo- rithm is evaluated in different conditions by changing the number of product types, .crane＇s dwell point and capacity in numerous simulations. The result shows that the crane＇s expected travel time can be reduced by the algorithm in different conditions.