Research On Pedestrian Evacuation Model Of Integrated Transportation Hub Based On Swarm Intelligence Algorithm

The integrated transportation hub(Integrated Transportation Hub,ITH)is crucial for passenger transportation and freight transportation,and it is closely related to the people's livelihood and economic development in the region.About the management of pedestrian evacuation within the hub,it has very practical significance to design reasonable evacuation models,which can provide path planning for the crowd and decision-making assistance for managers,so that reduces the investment for evacuation and improves ITH's operational efficiency and quality.In this paper,Chongqing North Station ITH is taken as the research object,meanwhile,evacuation path planning and evacuation simulation are regarded as the main research contents.Based on two-dimensional dynamic network,the multi-ant colony dual pheromone algorithm(Multi-Ant Colony Dual Pheromone Algorithm,MACDPA)for transfer evacuation path planning and dynamic partitioned ant colony algorithm(Dynamic Partitioned Ant Colony Optimization,DPACO)for emergency evacuation path planning are proposed respectively.Moreover,considering the schemes of two path planning algorithms as macro guidance,a pedestrian evacuation simulation model based on Cellular Automaton-Improved Artificial Fish School Algorithm(Cellular Automaton-Improved Artificial Fish School Algorithm,CA-IAFSA)was raised based on two-dimensional grid.The main works of the paper are as follows:(1)Research of transfer evacuation route planning.In order to solve the single-source multi-destination path planning problem that crowds evacuate from the railway exit to the target stations under conventional scene,the evacuees are divided into multiple ant colonies according to different transfer intentions.The MACDPA path planning algorithm with multi-ant colony planning and dual pheromone feedback strategy is proposed by adopting the population pheromone tribal pheromone into ACO.Experimental results show that: compared with ACO and HMERP-ACO algorithm under the same experimental conditions,the minimum network clearing time obtained by MACDPA algorithm is reduced by 32.6% and 5.2%,and the minimum cumulative congestion coefficient is reduced by 38.8% and 5.5%,the per capita evacuation time is reduced by 18.8% and 3.3%.As a result,the dual information cooperation mechanism of MACDPA can effectively balance the path load,make the search process jump out of local optimum,and increase the diversity of Pareto solution set while expanding the search range.(2)Research of emergency evacuation route planning.Under the disaster emergency scenario,in view of the fact that the MACDPA algorithm cannot solve the multi-source and multi-destination path planning problem of pedestrian evacuating from the initial position to the exit,considering the initial distribution of disaster sources and pedestrians for dynamic partitioning,the DPACO path planning algorithm is put forward based on the improved ant colony algorithm,which divides the path planning process into a prior distribution phase and a dynamic adjustment phase.The experimental results show that: when the number of evacuees is 500 and 5000 respectively,the minimum network emptying time obtained by DPACO algorithm is reduced by 10.2% and 20.67% than HACO',and the minimum cumulative congestion coefficient is reduced by 13.77% and 26.52% than HACO's.Compared with HACO algorithm,the DPACO algorithm can preferentially assign shorter paths to pedestrian,and reduce the evacuation time and cumulative congestion coefficient while ensuring that the evacuation time per person and the distance per person are small.(3)Research of pedestrian evacuation simulation.Based on the premise of the evacuation path planning which is obtained by the algorithms proposed in research content(1)and(2),AFSA algorithm is improved to comply pedestrian's moving on the top layer,in addition,CA model is applied on the bottom layer to resolve position conflict.Hence,a micro-evacuation simulation model based on CA-IAFSA algorithm is proposed to realize the combination of the macro path planning scheme and the micro evacuation motion simulation.The experimental results show that : the CA-IAFSA model can overcome the defect that the original AFSA is easy to fall into the local optimum in the evacuation environment with obstacle.When the number of evacuees is 500,with the minimum network emptying time evacuation scheme obtained by the DPACO algorithm,the CA-IAFSA model reduces the time complexity by 43.6% compared with the ICA model.Moreover,it has better simulation effect than CA in characterizing people's orientation,obstacle avoidance,follow-up behaviors,and can simulate the evacuation process in the ITH more realistically.(4)Design of pedestrians' evacuation system for ITH.According to the research results of this paper,from the aspects of system architecture,functional modules,database,etc.,the pedestrians' evacuation system for ITH is analyzed and designed,so as to construct a platform framework for evacuation decision support for ITH's managers.