Research On Optimization Route Decision Of Multi-modal Urban Transportation Network Based On Markov Decision Process

The phenomenon of separation between occupation and residence has been increasing with the development of urban economy,which has led to the increase of travel demand.In large and medium-sized cities,rail transit has become the primary mean of urban travel due to its large capacity,fast speed and high punctuality rate.Therefore,rail transit,bus,taxi,bicycle and walking constitute a multi-modal urban transportation network and can provide more travel choices for travelers.However,the multi-modal urban transportation network is also affected by uncertainties such as traffic flow and emergencies which may change the route decision of multi-modal transportation network.Based on this background,the travel decision problem of multi-modal urban transportation network under uncertain conditions are studied in this thesis.In order to find a solution to this problem,the relevant theoretical knowledge of route decision planning is introduced,the travel characteristics of multi-modal urban traffic is analyzed,and a multi-modal urban transportation super-network modal is proposed in this work.Since the Markov Decision Process has random characteristics,different states of things can be considered when making decisions.Therefore,a Markov-based method is used to solve the route decision problem.By comprehensively analyzing the factors that affect traveler's travel choice,a route decision model and an algorithm based on Markov Decision Process are proposed,and the feasibility of the proposed model and algorithm is verified by a real-world case.The main work of this paper is as follows:Firstly,since the travel time may be affected by uncertain factors such as traffic flow in transportation network,we divide different operational status into different states.Considering these different states,a route decision model with the target of minimum travel time is established and Markov Decision Process is used to solve the problem.Besides,according to the actual operation of the multi-modal urban transportation network,the proposed model considers the walking time and waiting time generated by the transfer between different modes.Secondly,for further improve the practicability of the model,this thesis analyzes the factors that influence travel decisions,establishes a route decision model with the target of minimum generalized travel cost which contains the travel time and travel cost,and designs the algorithm.Then,by using linear weighting method,we continue to explore the influence of these two factors on travel choices of different travel groups.Finally,a real-world case in Beijing is set as an example in order to verify the effectiveness of the route decision model and algorithm based on Markov Decision Process.The results show that there exist differences in travel decision results under different travel states.Markov-based methods can provide a more robust route decision scheme for travelers,and the proposed models and algorithms are effective.The thesis focuses on the theoretical study of route decision model in multi-modal urban transportation network,involving urban rail transit,regular bus,walking,bicycle and taxi.By considering different traffic states,we propose a new space for discussion in this field.The model and algorithm have practicability which can provide a theoretical method reference for the development of Intelligence Transportation System.