时变排队模型下的高校食堂排队预测研究

Queue Prediction in University Dining Halls Using a Time-Varying Queueing Model

摘要: 针对高校食堂高峰拥堵、客流监管难等问题，本文首先将就餐窗口抽象为多服务台排队系统，搭建排队预测与推荐框架；其次采用$M_t/G/c\left(t\right)$ 离散事件仿真描述早、中、晚主要用餐高峰，并在非高峰时段采用低峰客流函数与$M/G/c$ 近似模型进行补充预测，覆盖高峰期并依据实际约束生成模拟数据；最后系统可输出多项运行指标与可视化图表，刻画设定仿真场景下的食堂客流特征，为就餐、窗口调度及备餐工作提供参考。

Abstract: To address peak-hour congestion and difficulties in monitoring diner flows in university dining halls, this study models each service window as a multi-server queue and develops an integrated queue prediction and recommendation framework. A time-varying $M_t/G/c\left(t\right)$ discrete-event simulation model is used to characterize the major breakfast, lunch, and dinner peaks. For off-peak periods, a low-arrival-rate function combined with an $M/G/c$ approximation is employed, and simulation data are generated under practical operational constraints. The framework produces operational indicators and visualizations to characterize diner-flow patterns under the specified simulation scenarios, providing analytical support for dining choices, service-window scheduling, and meal preparation.