Study On Epidemic Spreading And Vaccine Allocation Strategy Based On Multilayer Networks

Complex networks provide powerful analytical tools for describing contact relationships in human society,particularly the multilayer networks,which can provide a comprehensive representation of the diverse types of social relationships that exist.Multilayer networks can be categorized into two types: interconnected networks and multiplex networks.The former is suitable for modeling multiple regions characterized by interconnected relationships,while the latter is better suited for representing multiple types of relationships within a particular area.In fact,the complex nature of epidemics are influenced by multiple social relationships,including family,school,workplace connections,in addition to external regions,multilayer networks are an ideal tool for studying epidemics.In light of the ongoing novel coronavirus outbreak,we have implemented two types of multilayer networks to model epidemics and study immune strategies.Our research work consists of the following two parts:(1)We constructed an interconnected multilayer network,consisting of two randominterconnected networks with varying average degrees to represent two regions with different internal contact frequencies,and studied the impact of vaccine distribution in different network layers on the entire network spreading range.Simulation and analysis results indicate that in order to achieve optimal control,the following vaccine allocation strategy should be followed:(1)rank network layers from small to large based on the threshold number of vaccines required by each network layer,(2)compare the total number of vaccines with the threshold number of each network layer,if the total number of vaccines is sufficient to suppress the spread of epidemics in l network layers,allocate vaccines to l + 1 network layers most likely to be suppressed,and(3)distribute 80% of the threshold number of vaccines required for each network layer until l + 1 network layers have been vaccinated.Furthermore,the effectiveness of this vaccine allocation strategy was validated by applying it to four real-world interconnected multilayer networks.(2)We constructed multiplex networks based on macro statistical data,where each layer represents different contact networks of the same individual within the same region.We investigated the data-driven process of constructing multiplex networks and the impact of different initial infection ages on the spreading of other ages and the spread of epidemics across the entire network.The constructed contact matrix and fitted parameters indicate that our network successfully describes the contact characteristics of different age groups in different network layers with higher Pearson coefficients compared to empirical data.This finding demonstrates that our contact matrix is in good agreement with empirical data.More specifically,our simulation and analyses revealed that the age of the initial infected person has no effect on the final spreading range and the order of the peak infection of the network.The proportion of infected individuals reaches the peak in each age group in the following order: ”school-age children and adolescents → middle-aged and young → young children → elderly”.However,the initial infection of school-age children and adolescents has a significant promoting effect on the peak infection arrival time of other ages.