基于SARIMA-BP神经网络对消防救援人员安排的预测

Prediction of Fire Rescue Personnel Arrangement Based on SARIMA-BP Neural Network

针对某地消防救援人员的安排,将消防救援出警次数分成两段,以2016年1月~2020年12月数据用于构建模型,2021年1~12月数据用检验模型,最终用于预测2022年1~12月.综合运用SARIMA模型、ARIMA模型等方法,分别对1~9月出警次数进行拟合并建立相应模型,并对10~12月出警次数进行预测.借助交通网络图结合Floyd算法,计算得出没有直接路径的各区域之间的最短距离.通过评价模型的准确性和稳定性,由此预测2022年各月份的消防救援出警次数,综合考虑区域之间的距离及工作均衡度,可以知道D、P、M三个区域是依次设立新消防站的最佳地点.

According to the arrangement of fire rescue personnel in a certain place,the times of fire rescue are divided into two sections.With the data from January 2016 to December 2020,the model is built,and with the data from January to December 2020,the model is tested,and finally the model is used to forecast from January to December 2022.In this paper,SARIMA model,ARIMA model and other methods are comprehensively used,respectively,to combine and establish corresponding models for the number of police calls from January to September,and to forecast the number of police calls from October to December.With the help of traffic network diagram combined with Floyd algorithm,the shortest distance between regions without direct path is calculated.By evaluating the accuracy and stability of the model,the number of fire rescue calls in each month in 2022 is predicted.Considering the distance between regions and the work balance,we can know that D,P and M are the best places to set up new fire stations in turn.