Abstract: To improve the management of the distribution center and reduce the queue time of delivery vehicles waiting for loading and unloading operations in the logistics park,the platform scheduling problem in the distribution center was studied.By considering the platform-vehicle compatibility constraint and the time window constraint of each vehicle,a mathematical model of the problem was established with total weighted earliness and tardiness penalties as the objectives.A learning hybrid differential evolution algorithm was developed to solve the problem.In this algorithm,an encoding and decoding method was designed according to the characteristics of the problem,a learning operator  selection mechanism was used to select crossover operators and mutation operators online,and the variable neighborhood search algorithm was adopted as the local search method to enhance the search ability of the algorithm.Then,an orthogonal test was carried out to determine the parameters of the algorithm.Comparative experiments were carried out which showed that the algorithm could effectively solve the platform scheduling problem and help the distribution center to better manage the platform scheduling.

Key words: platform scheduling, learning hybrid differential evolution algorithm, unrelated parallel machines, total weighted earliness and tardiness penalties

摘要： 为提高配送中心的管理水平,减少物流园区中等待作业的配送车辆的排队时间,研究了配送中心中的月台调度问题。首先,考虑月台—车辆兼容性约束和车辆作业时间窗约束,建立了以最小化总加权提前、拖后惩罚为目标的数学模型;然后,提出一种学习型混合差分进化算法,根据问题特征设计了月台调度问题的编解码方法,设计了一种学习型算子选择机制为算法在线选择交叉、变异算子,采用变邻域搜索算法作为局部搜索算法增强算法的搜索能力;最后,通过正交试验,确定了算法参数水平,进行对比实验,证明了所提模型和算法能够有效求解月台调度问题,从而帮助配送中心更好地进行月台调度管理。

关键词: 月台调度, 学习型混合差分进化算法, 不相关并行机, 总加权提前、拖后惩罚