Open AccessPassenger’s routes planning in stochastic common-lines’ multi-modal transportation network through integrating Genetic Algorithm and Monte Carlo simulation
Author(s) -
Yunfeng Peng,
Zhiyao Mo,
Lina Song
Publication year - 2021
Publication title -
archives of transport
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.309
H-Index - 14
eISSN - 2300-8830
pISSN - 0866-9546
DOI - 10.5604/01.3001.0015.0123
Subject(s) - genetic algorithm , population , monte carlo method , dwell time , computer science , flow network , duration (music) , mathematical optimization , transport engineering , operations research , simulation , engineering , mathematics , medicine , art , clinical psychology , statistics , demography , literature , sociology
In the urban transportation network, most passengers choose public transportation to travel. However, bad weather, accidents, traffic jams and other factors lead to uncertainty in transportation network. Besides, transport vehicles running on the same segments of routes and belonging to different modes or routes make the transportation network more complicated. In order to improve the efficiency of passenger’s travel, this paper aim to introducing an approach for optimizing passenger travel routes. This approach takes the travel cost and the number of transfers as constraints to finding the shortest total travel duration of passenger in urban transporta-tion network. The running duration and dwell duration of the vehicles in the network are uncertain, and the vehicles are running according to the timetables. As transportation modes, bus, rail transit and walk are consid-ered. In terms of methodological contribution, this paper combines Genetic Algorithm (GA) and Monte Carlo simulation to deal with optimization problem under stochastic conditions. This paper uses Monte Carlo simula-tion to simulate the running duration and dwell time of vehicles in different scenarios to deal with the uncertainty of the network. The shortest path of passenger’s travel is solved by GA. Two kinds of population management strategies including single population management strategy and multiple population management strategy are designed to guide the solution population evolving process. The two kinds of population management strategies of GA are tested in numerical example. The satisfactory convergence performance and efficiency of the model and algorithm is verified by the numerical example. The numerical example also demonstrated that the multiple population management strategy of GA can get better results in a shorter CPU time. At the same time, the influ-ences of some significant variables on solution are performed. This paper is able to provide a scientific quantita-tive support to the path scheme selection under the influence of common-lines and timetables of different modes of transportation in stochastic urban multimodal transportation network.