Research On COVID-19 Epidemic Prevention And Control Strategies Based On Deep Reinforcement Learnin

At present,the global COVID-19 is still severe.More and more countries have experienced the second,third,and even multiple outbreaks.The epidemic is still far from over,even if the vaccine is successfully developed and put on the market on a large scale.Inappropriate epidemic control strategies may bring catastrophic consequences.It is essential to maximize the epidemic restraining and to mitigate economic damage.However,the study on the optimal control strategy concerning both sides is rare,and no optimal model has been built.The existing general infectious disease dynamics model is the SusceptibleInfectious-Hospitalized-Recovered(SIHR)compartment model compartment model.SIHR model can effectively simulate the development of epidemics and make predictions.However,SIHR model does not consider the impact of government prevention and control measures on the development of the epidemic.To reflect the relationship between prevention and control and the economy in the model,this paper improved the SIHR model and established economic models affected by quarantine measures.On this basis,we use the deep reinforcement learning method to formulate a life-economic optimal prevention and control strategy.In this paper,SIHR model is expanded to simulate the epidemic’s spread concerning isolation rate.An economic model affected by epidemic isolation measures is established.The effective reproduction number and the eigenvalues at the equilibrium point are introduced as the indicators of controllability and stability of the model and verified the effectiveness of the SIHR model.Based on the Deep Q Network(DQN),one of the deep reinforcement learning(RL)methods,the blocking policy is studied to maximize the economic output under the premise of controlling the number of infections in different stages.Finally,we considered the impact of vaccines on the development of the epidemic and established the Susceptible-Infectious-Hospitalized-Recovered-Immune(SIHRM)model.In this part,the parameters of deep reinforcement learning are adjusted accordingly,and the impact of vaccination rate and virus mutation on the development of the epidemic,and the change corresponding economic-life optimal epidemic prevention and control strategy are studied.The study demonstrates the following important results:(1)optimal policies may differ in various countries depending on disease spread and anti-economic risk ability.(2)The epidemic prevention and control strategy that focuses too much on the economy can reduce economic losses in the short term and help economically fragile countries avoid serious economic crises.But the epidemic is more likely to break out again.(3)The deep RL can be applied to different infectious disease models and develop corresponding optimal strategies.(4)The impact of vaccination rate on the development of the epidemic is the most significant.With the increase in vaccination rate and the number of immune individuals in the environment,the government can take more relaxed isolation measures to quickly restore the economy.However,the government needs to be vigilant against the mutation of the virus to prevent the outbreak from spreading again.This thesis concludes that a policy that can both control the epidemic and restoring economy is given using the method of deep reinforcement learning training.And recommend that the government cannot completely relax the control of the epidemic.While the government is strengthening inspection measures,the public must continue to receive new vaccines.