Path Planning For Guided Passengers During Evacuation In Subway Stations Based On Multiobjective Optimization

Most subways are built underground,with relatively enclosed spaces,large daily passenger flows and many directions of flow,which can easily lead to casualties in the event of an accident.One of the effective ways to improve evacuation capacity is through active guidance and intervention in the direction of controlled passenger movement.However,there is a lack of data on passenger evacuation in subway stations,and research on the intervention of guidance information in evacuation is not yet in-depth.This paper focuses on the evacuation of passengers subject to the intervention of guidance information under clearance situations such as train failure and massive power failure in subway stations.A multi-objective optimization method is proposed for the evacuation paths of passengers guided by guides and applets respectively.Relevant analysis on the evacuation effect under the action of guidance information is carried out,with a view to timely evacuating the passenger flow in stations and improving the level of emergency services in subway stations.The main contents of this paper are summarized as follows.(1)Simulation model validation.The Mass Motion simulation model is built based on the actual scenario data of the subway station,and the validity of the model is verified.The validity of the social force model for describing passenger movement behaviour is verified by comparing the field observation data on the walking time of exiting passengers with the simulation data,and by comparing the field observation data on the number of passengers passing through the gates with the simulation data.The validity of the minimum cost model for modeling passenger path selection behavior is verified.The effectiveness of the social force model and the least cost model in driving passengers through the nodes is verified by comparing the field observations of passengers’ passage time at the staircase with the simulation data using independent sample T-tests and non-parametric-independent sample tests.(2)Development of a multi-objective optimization method for passenger evacuation paths under guides.A scheme for the allocation of the initial number and location of guides in subway stations is established based on the Gaussian mixture model and the cost function method.The model takes into full consideration the distribution of passengers in the station,in order to achieve the objectives of increasing the induced range,reducing the waste of human resources and improving the evacuation efficiency.Furthermore,a multi-objective optimization model with the shortest evacuation time and the smallest congestion cost is established for determining the evacuation routes to avoid evacuation bottlenecks in advance.Simulation experiments under two evacuation scenarios with/without guides based on the social force model and the minimum cost model and the guide induction scheme are conducted.The changes in passenger evacuation time,passenger density and evacuation path selection under guides are analyzed,and the effectiveness of the multi-objective optimization model for passenger evacuation paths under guides proposed in this paper is verified.(3)Establishment of a multi-objective optimization method for passenger evacuation paths considering node efficiency under an applet intervention.Considering the influence of passenger passage capacity to the nodes such as gates and stairs/escalators on passenger evacuation efficiency,a multi-objective optimization model with the objective of minimizing the maximum evacuation time of heterogeneous passengers in different segmentation areas is established to achieve the local optimization on the basis of global optimization of evacuation paths.Considering the influence of passenger heterogeneity,the atomic orbital search algorithm is used to improve the BP neural network,thereby a novel node time prediction model is proposed.The effectiveness of the node time prediction model is verified by analyzing the prediction results.The effectiveness of the multi-objective optimization method for evacuation paths considering node efficiency is verified by analyzing the changes in passenger evacuation time and density.The influence of various factors on passenger evacuation time is identified through sensitivity analysis.A We Chat applet is designed to deliver evacuation information to passengers to achieve the goal of efficient guidance.