A Study On The Location And Distribution Of Epidemic Prevention Supplies Storage Centers In The Context Of The COVID-19

Today,many outbreaks of public health events,represented by infectious disease epidemics,which threat people’s health,and affect economic development.Whether from an economic or social perspective,public health outbreaks can cause tremendous harm to humanity.In the face of these outbreaks of infectious diseases and epidemics,it is crucial to achieve accurate forecasting of epidemic trends and efficient strategies for the distribution of emergency supplies.This paper uses the theories of infectious disease dynamics,deep learning and logistics network optimization as the foundations to construct a theoretical system for implementing forecasts of epidemics and designing logistics networks based on the prediction results,using the 2019 outbreak of Covid-19 in China as the background,and applies the theoretical system to S city in China,realizing an example application based on real data.The core research of this paper has the following parts below: Firstly,based on the theory of infectious disease dynamics,the basic SEIR model is extended to take into account the situation that recovered patients may be re-infected into the model,in order to achieve a more appropriate prediction of the epidemic;Secondly,using the LSTM model in deep learning,the model is constructed and trained based on the real data of the new crown pneumonia in China;Finally,build an example application based on the theoretical system described above was carried out with the city of S city as the subject of the study.According to this study,some conclusions are clearly show as below:1)the infectious disease dynamics model focuses more on the impact of internal changes in the population system on the prediction trend by dividing the population into multiple "compartments" internally,and thus it is difficult to acquire the impact of external interventions such as epidemic prevention policies during the new crown period on the trend of the epidemic.2)the univariate LSTM neural network model was found to be more effective in predicting confirmed patients in this paper,probably because the increase or decrease in confirmed cases itself reflects the influence of external epidemic prevention and control policies on the epidemic trend,and the inclusion of too many explanatory variables may not yield satisfactory results.3)in the sensitivity analysis of the site-allocation model This is because,in epidemic prevention activities,if the number of alternative sites is higher,then the time to serve neighboring relief sites will be shorter and the corresponding service cost will become lower as the distribution distance decreases,which is in line with the reality.