Application Research On Ticket Revenue Management Of High Speed Railway Based On Data-driven

In recent years,with the rapid expansion of the scale of China’s high-speed railway network,high-speed railway has become one of the important transportation modes for people to travel,and has made great contributions to China’s economic development and resource allocation.However,under the unprecedented prosperity of high-speed railway,there are also some problems.From a macro point of view,compared with ordinary high-speed railway,high-speed railway needs to invest more funds in infrastructure construction and daily operation,which has caused a huge burden to railway transportation enterprises;From the perspective of daily operation,due to the uneven distribution of passenger flow demand in time and space and the characteristics that transportation products are produced and consumed at the same time and cannot be stored,there are two contradictory phenomena of insufficient transport capacity and empty transport capacity in the whole country.At present,the ticket price structure of high-speed railway is single,which can not adjust passenger flow and maximize ticket revenue.Based on this,this paper takes the high-speed railway ticket revenue management as the research object,combined with the characteristics of China’s high-speed railway transportation organization,studies the high-speed railway ticket application revenue management theory from four aspects: passenger flow characteristic analysis,short-term passenger flow prediction,ticket allocation and dynamic pricing,and studies a set of theoretical methods of high-speed railway ticket application revenue management driven by big data,In order to enhance the competitiveness of high-speed railway transportation in the field of transportation and increase the ticket income of high-speed railway.The specific contents include:(1)The analysis method of passenger flow fluctuation characteristics of high-speed railway under big data is studied.By analyzing the characteristics and influencing factors of passenger flow fluctuation of high-speed railway,according to the characteristics of passenger flow fluctuation,this paper puts forward the feature extraction method of passenger flow time fluctuation based on Prophet model and the analysis method of influencing factors of passenger flow fluctuation based on XGboost algorithm.The former can decompose the time characteristics of each time distribution type from the time series,and the latter can quantify the impact of each feature on passenger flow fluctuation from the feature attribute matrix.Finally,taking the historical passenger flow demand data of an OD interval for nearly 15 months as the object,a case study is carried out to verify the feasibility and effectiveness of the model.(2)The short-term passenger flow prediction method of high-speed railway based on neural network combination model is studied.Using the model algorithm of deep learning,the improved model of convolutional neural network and cyclic neural network in time series prediction is combined,and the attention mechanism is introduced to construct the short-term passenger flow prediction model of high-speed railway based on TCN-LSTM-Attention.At the same time,through the results of passenger flow characteristic analysis,a reasonable and effective passenger flow characteristic matrix is created to train and predict the model.Finally,taking the historical passenger flow demand data of an OD interval for nearly 15 months as the object,the passenger flow prediction and evaluation are carried out,and compared with other methods.The results show that the prediction model proposed in this paper has better prediction effect.(3)A multi train dynamic ticket allocation method considering passenger demand orientation is studied.Firstly,by analyzing and subdividing the passenger demand,it is proposed that the passenger demand for high-speed railway is mainly divided into two parts:transportation demand and preference demand.Then,with the goal of meeting the passenger transportation demand as much as possible and meeting the passenger’s preference demand for products,and taking the seat capacity of each train in each section as the constraint,a multi train dynamic ticket allocation model considering the passenger demand-oriented random demand is constructed,and the model is solved by combining the branch and bound method and simulated annealing algorithm.Finally,a case study of dozens of trains in a line section is carried out to verify the effectiveness of the model algorithm.(4)The dynamic pricing strategy of high-speed railway ticket based on reinforcement learning is studied.Firstly,the characteristics of daily pricing behavior of high-speed railway during the ticket pre-sale period are analyzed,and then according to its characteristics,a Markov decision process model is constructed with the goal of adjusting transportation demand and maximizing ticket revenue,and a reinforcement learning algorithm is designed to solve the optimal strategy in each stage.Finally,the effectiveness of the model is verified through the comparison of large and small cases in the light and peak seasons of passenger demand.The results show that the dynamic pricing strategy can effectively adjust the balance of supply and demand,and achieve the purpose of improving ticket revenue while reasonably optimizing resource allocation.