Study On Travel Path Planning Based On Epidemic Related Regional Risk Assessment

In 2020,the COVID-19 epidemic began to spread widely.Due to the latent and asymptomatic nature of the virus,it has a great impact on people’s transportation.People must consider the travel risks before traveling,and try to achieve healthy and safe travel.Although COVID-19 is no longer as rampant as before,and the whole epidemic situation is gradually developing in a better direction,there is no guarantee that a more virulent virus will mutate or similar public health and safety events will occur again in the future.During the recovery period of the epidemic,the willingness of individuals to travel increased significantly,but according to the relevant travel navigation software,the planned route will still pass through the epidemic risk associated areas,and there is still a large risk of infection in the travel process.In order to solve the above problems and further improve the safety and travel efficiency of residents during the epidemic or other public health events,his paper studies the travel path planning model and algorithm based on the risk assessment of epidemic related areas.The specific research contents are as follows:First of all,this paper describes in detail the issues related to risk assessment based on epidemic related regions,expounds the theory of spatial interaction,and analyzes the four main modes of transmission of the COVID-19 epidemic.The risk assessment model of the associated area is established with the aid of the migration-type transmission model.The intensity of the interaction between the associated areas is estimated according to the principle of the gravity model.The improved one-dimensional steady-state water quality migration model is used to assess the epidemic spread risk of a single point source,and the epidemic risk of the surrounding roads is quantified.Then,a risk aversion path planning model based on reinforcement learning is designed,and the problem description and model establishment of risk aversion path planning behavior under the epidemic situation are carried out.The state and action space of the agent in the reinforcement learning process are designed in a "point-to-point" way,and the reward function is designed according to the location of the agent,the location of the risk point and the starting and ending positions.The total length of the path and the distance from the epidemic risk area are taken as objective functions to solve the problem.A method for generating impedance matrix of road network based on SUMO simulation is proposed,which solves the problem of low efficiency in building road network model manually.Secondly,a restricted search reinforcement learning(RRL-APF)algorithm is proposed to improve the artificial potential field.The restricted search method is used to optimize the Q table initialization strategy and improve the Q value of favorable actions.The improved artificial potential field method is designed to optimize the action selection strategy of the agent.A dynamic greedy strategy method is designed to adjust the greedy coefficient of each reinforcement learning,effectively balancing the exploration and utilization of the agent,Improve convergence speed and stability.The modeling and solving processes for global path planning and dynamic path re planning considering the epidemic situation were designed.Finally,in order to verify the effectiveness of the algorithm,this paper takes the surrounding area of Xinfadi,Fengtai District,Beijing as an example,sets different travel scenarios and learning parameters,and compares the RRL-APF algorithm proposed in this paper with Q-learning algorithm,Sarsa algorithm and strong chemical habit algorithm based on artificial potential field(RLAPF).The results show that under different travel scenarios,the convergence speed of the algorithm proposed in this paper is significantly improved compared with the other three algorithms,and the travel distance can be reduced to shorter while ensuring safety.For the dynamic path re planning problem considering changes in the epidemic situation,the algorithm process of dynamic path re planning was validated by adding new risk areas,and the risks of the global route and the re planned route were compared.This paper can provide people with a path plan that takes into account both safety and travel distance during the epidemic,and can provide a reference for travel planning after the recurrence of related public health events or emergencies in the future.