Research On The Spread Risk Of Epidemic Assessment Model Based On Subway Passenger Flow

Since the beginning of this century,there have been frequent outbreaks of complex epidemics,mainly transmitted in cities through contact between citizens.In order to control the transmission,this paper first divides the traffic sub-areas according to the commuter flow,and then carries out the travel restrictions in different sizes of the regions and intervals by taking the sub-areas as units.This paper aims to study the regional risk rating algorithm based on traffic passenger flow,control the cross-regional transmission of epidemic diseases by controlling the scale of travel restrictions,and provide quantitative reference for regional prevention and control.Firstly,this paper combines the network centrality,graph convolutional neural network and the passenger flow data of urban traffic subareas,and proposes the regional propagation risk rating algorithm SRR.The algorithm uses time-varying subinterval passenger flows,and considers the neighborhood connectivity of regions,the hub role of regions and the iterative steady-state distribution of network risk transmission.Due to the time-varying characteristics of passenger flow,this paper proposes a time-sharing propagation risk rating algorithm HSRR.Then,combined with the ranking iteration results of HSRR algorithm,this paper proposes a region propagation risk ranking algorithm GCN-GPR based on graph convolution network.In this paper,SRR,HSRR and GCNGPR are all iterative algorithms,that is,the propagation risk index of the region is iteratively transmitted by the adjacent region.According to the transmission risk rating algorithm,travel restrictions of a certain scale are carried out to reduce the loss of passenger flow caused by control transmission as much as possible.Second,this paper constructed a spatial-temporal variation of citizen contact and infection network,combined with individual health status model and multi-scenario citizen contact network model,and the transmission was based on the travel trajectory of each citizen rather than the transmission rate of group contact.In order to verify the effectiveness of the regional risk rating algorithm,this paper first built an inter-regional simulation platform of Shenzhen City based on real passenger flow data.Citizens traveled on the shortest route in history and contacted different citizens within the region,at traffic stations and between regions.Then the scene contact network with modified parameters was used to model the epidemic transmission process in the platform.Then,the interregional traffic restriction was simulated according to the risk rating index of the urban region,and the 28-day epidemic transmission simulation experiment was conducted according to the inter-regional traffic restriction simulation of different sizes in the simulation platform.Finally,it is found that the passenger flow loss of GCN-GPR algorithm is 67.1%,which is better than HSRR(1%),SRR(3.7%)and traditional network centrality algorithm(9.40%DC,13.20%BC,12.00%PG).Is the algorithm with the lowest passenger flow loss among all the algorithms,and the regional centrality algorithm based on time-sharing passenger flow is better than the regional centrality algorithm based on daily passenger flow.