Research On Revenue Management For Dedicated Passenger Line Based On Passenger Choice Behavior

To meet the growing inter-city passenger transport demand, our country is further extending and expending the coverage of fast passenger transport network at the same time with constructing the "four-horizontal-four-lengthwise" dedicated passenger lines. Construction and operation of dedicated passenger lines greatly ease the railway passenger transport capacity bottlenecks, however, inevitably intensify competition among various transportation modes in domestic passenger market, that makes the marketization trend of dedicated passenger line operation further embody. The marketization operating experience for many years of railway passenger transport companies of the United States, France, Britain, Germany and other countries shows that revenue management is an effective method to optimize structure and allocation of resources and enhance the competitive capacity and operating income. Therefore, the research on revenue management of dedicated passenger line, especially the one based on passenger choice behavior, rather than only from the perspective of dedicated passenger line, which not only reflects the thought of passenger oriented, but also belongs to the hot spots, has important theoretical and practical significance, and provides a reference and a new perspective for dedicated passenger line to implement revenue management.In Chapter 1, the background was analyzed from the two aspects of theory and planning practice, the feasibility and necessity of implementing revenue management and the significance of studying revenue management for dedicated passenger line were elaborated, the problems to be studied in the future were put forward based on the systematic analysis of the application and research status of railway revenue management.The research achievements of revenue management and passenger choice behavior were summarized in Chapter 2, thus to effectively grasp the development history, basic connotation, research contents and foundation models of revenue management, to make clear the research status and frontier dynamic of dynamic pricing and inventory control, and to understand the underlying theory and common types of consumer choice behavior in revenue management, which provides the basis and lay the foundation for expansion of the subsequent content.In Chapter 3, a mixed regression model was constructed to explore the relation of passengers’overall evaluation on dedicated line product to their evaluation on safety, comfort, speed, frequency, punctuality, price, convenience seven attributes. The regression coefficients were estimated by EM algorithm, the probability of passenger belonging to different types was calculated according to the Bayesian statistical theory, and the munber of types was determined by Bayesian Information Criterion and Akaike Information Criterion. Combined with the passenger survey data on Wuhan-Guangzhou passenger dedicated railway, the passenger market was subdivided into four types, and the social and economic characteristics and travel demand characteristics of different types of passengers were analyzed by correlation analysis.In Chapter 4, Making use of passenger reservation price to describe the passenger choice behavior, assuming the reservation price of passengers that order the same path to be independent and identically distributed, combined with the known fare sets and booking passengers’arrival probabilities, a dynamic programming model was established according to Bellman optimization principle, which allows to optimize the expected revenue through dynamic adjusting the ticket price. Further, the optimal ticket price was proved to increase with its marginal expected revenue and the optimal strategies of the dedicated passenger line consisted of one and two legs to be characterized of threshold respectively.The revenue optimization problem ofdedicated passenger line with two fare levels and multiple legs was studied in Chapter 5. Assuming passenger’buy-up behavior exsists and its probability is known, combined with the given ticket prices and demand distribution, a constrained nonlinear integer programming model was established by taking the nested booking limits of each kind of tickets of each itinerary as decision variables. The model was solved through obtaining the solution generation point, generating initial particle swarm, calculating fitness values, updating particle position four steps.In Chapter 6, as for the revenue optimization problem ofdedicated passenger line with multiple legs, multiple trains and multiple fare levels, adopting passenger preference order to depict choice behavior, combined with the given ticket and passenger category features, a dynamic programming model was constructed by taking ticket control strategies in each booking period as decision variables. In view of its large scale, the deterministic linear programming model was built to approximate it and solved by using the column generation algorithm and genetic simulated annealing algorithm. Then, the original dynamic programming model was heuristicly decomposed by making use of the optimal dual solutions. Finally the expected revenue of the dedicated passenger line was optimized approximately.