Design And Strategy Research Of Simulation Platform For Passenger Flow Service Optimization In Smart Towns

Smart towns have become an important trend in future urban development,because human beings are the core driving force of urban development,and their needs and behaviors affect the development of cities.With the increase of urban population,the problem of passenger flow service also occurs,so how to combine passenger flow service with intelligent design of smart town has become an important research content in the field of information technology and passenger flow service.The pedestrian simulation platform is a tool that can simulate pedestrian behavior and movement,which can support the research of passenger flow service optimization problems.However,the existing simulation platform has certain functional shortcomings,which cannot meet the needs of smart town passenger flow service optimization scenarios.Therefore,this thesis aims to realize a simulation platform for operation managers and researchers,and to study the strategy algorithm for passenger flow service optimization in smart towns through this simulation platform.1)According to the needs of the smart town passenger flow service optimization scenario,the simulation platform design is carried out.Simulation platform design includes three aspects:simulation platform overall architecture design,simulation platform function analysis and design,and simulation platform back-end function analysis and design.According to the demand analysis and simulation platform design scheme,realize the specific simulation platform function,and explain its implementation principle.This includes simulation platform simulation aspects,simulation platform strategy service aspects,and simulation platform information management service aspects.The simulation platform simulation aspect implements the basic simulation functions of the simulation platform,including environmental modeling,pedestrian movement modeling,visual display and other agent settings.In terms of simulation platform strategy service,the strategy research related functions of the simulation platform are realized,including the application and training of strategy models,agent interaction,simulation process control,data services and deployment and application functions of policy models.In terms of simulation platform information management service,the simulation platform user-oriented simulation information initialization configuration service is realized,and users can complete the pre-preparation for creating simulations.Finally,the simulation platform designed in this thesis is verified by functional testing.2)Aiming at the congestion mitigation scenario in the optimization of passenger flow service in smart towns,this thesis proposes a congestion mitigation algorithm based on DDQN algorithm.In this thesis,the cascade of multi-time pedestrian distribution states is used as the status input,and the issuance of coupons is used as the drainage measure.In the policy learning,the agent adopts the exploration scheme guided by the dynamic optimization strategy proposed in this thesis.Finally,simulation experiments are verified in the simulation platform designed in this thesis.Experimental results show that the algorithm proposed in this thesis can effectively alleviate congestion,and the convergence speed of the strategy is faster,and the convergence peak is also improved.