A Study on Prediction and Scheduler for Foot Traffic of University Canteens Based on Gray Prediction and BP Neural Network

Abstract

In the current context of epidemic prevention and control, the resumption of university education places greater emphasis on the prediction and regulation of foot traffic in public areas. Using the university canteens as an example, predicting the foot traffic during the dining period benefits the canteens staff's reasonable scheduling, reducing the potential virus transmission risk caused by the dense crowd, as well as providing time- sharing service for the distribution of the canteens foot traffic, reducing dining waste and practicing diligence and frugality, and helping to alleviate the crowded queue during the dining period. Based on field research on canteen patronage data, this paper combines gray theory, neural network technology research, and patronage information prediction research to adapt to the nonlinear characteristics of patronage, optimize the performance of the prediction model, and improve the accuracy of the prediction model. The results are used to determine patronage density and the number of canteen windows. This paper used factor analysis to build a refined foot traffic guidance and schedule according to each index and put it into the smart canteen website.