Self-organizing Voluntary Vaccination In Complex Networks

Infectious diseases have been a serious threat to human health and social economics.Understanding human voluntary vaccinating behaviors plays essential roles in eradicating a vaccine-preventable disease.Recently,the combination of complex network theory and epidemiology has become a major trend in epidemiological research.Based on existing studies,we propose and evaluate several vaccinating mechanisms.The main contribution of our paper is as follows:First,when considering whether or not to take vaccine,the rational individuals usually make decisions by weighing both the cost of vaccination and infection.However,individuals' vaccinating decisions can also be influenced by vaccinating behaviors of their social neighbors.In view of this,we consider both individual's historical vaccinating experiences and the impact of social influence,and propose a self-organization vaccinating mechanism to investigate human voluntary vaccinating behaviors.Through simulation on disease transmission and human vaccination behaviors in complex networks,we examine the combined effects of both individuals'social influence and historical vaccinating experiences on the final vaccine coverage under the different relative cost of vaccination and infection.The simulation results show that the impact of social influence-based behavior on public vaccine coverage is greater than that of rationality-based behavior,indicating the importance of social guidance in disease intervention and control.Second,Individuals' vaccinating decisions are influenced by both historical vaccinating experiences and social neighbors when making vaccinating decisions.But the extent to which an individual makes vaccinating decisions based on rationality is a rather confusing issue.In this paper,we propose a reinforcement learning-based mechanism to characterize human voluntary decision-making process with bounded rationality,where each individual can adaptively choose to make vaccinating decision based on historical experiences or social influence.Through simulations,we investigate the performance of decision-making mechanisms with/without reinforcement learning in terms of vaccine coverage,final epidemic size,average payoff,and vaccine effectiveness under the different relative cost of vaccination and infection.The simulation results show that reinforcement learning can improve vaccine effectiveness by balancing individuals' rationality and social influence.It demonstrates the importance of appropriately utilizing human bounded rationality for preventing disease epidemics.At last,the perception of epidemic severity can also affect individuals' vaccinating decisions.During the epidemic of a vaccine-preventable disease,individuals in social networks can perceive the risk of infection based on either global epidemic information announced by the public health authorities,or local epidemic information obtained from their social neighbors.After that,they can rationally decide whether or not to vaccinate by weighing the cost of vaccinating and infection.In this case,both the social network structure and the individual's risk perceptions strategies will influence the final vaccine coverage.Accordingly,we propose three types of static decision-making mechanisms,namely the Local—G1 mechanism,the Local-G1 mechanism,and the Global-G mechanism.Each mechanism simulates human vaccinating behaviors based on different local or global information.Furthermore,we propose an adaptive decision-making mechanism based on reinforcement learning,individuals can adaptively determine which information to use to estimate the severity of the epidemic based on their historical vaccinating experiences.Simulations results on three types of social networks show the impact of network structure,source of information,relative vaccination risk,and individual social connections on the final vaccine coverage and epidemic size.In summary,this paper has the following contributions:(1)We present a self-organizing vaccinating mechanism that integrates both individuals' rationality and social influence to investigate individuals' voluntary vaccinating behaviors,and find that social guidance plays an important role in disease intervention.(2)We present a reinforcement learning mechanism to characterize human voluntary decision-making process with bounded rationality and find that the importance of appropriately using human bounded rationality to prevent disease epidemics.(3)We present a voluntary vaccinating mechanism based on the perception of epidemic severity.Individuals can perceive the risk of infection based on either local or global epidemic information.we find that individuals based on epidemic information from second-order neighbors will overestimate the risk of infection,and high degree individuals will overestimate the risk of infection.The results and findings in this work can provide new insights into the design of incentive-based vaccinating policies and intervention strategies for seasonal epidemics.