TL;DR
This paper tackles the challenge of maximizing high-speed rail revenue by jointly optimizing group ticket discounts and dynamic pricing for individual passengers. They propose a dynamic programming model that dynamically controls group booking limits, showing it outperforms fixed limits and simple dynamic pricing, especially when individual demand is low.
What Problem Does This Paper Address?
High-speed rail operators face a classic revenue management dilemma: how to balance discounted group tickets with dynamic pricing for individual travelers. While group discounts can stimulate demand, selling too many can cannibalize higher-paying individual customers. Conversely, not enough group sales mean missed revenue opportunities. The core problem is finding the optimal "symmetry point" where the expected revenue from accepting group bookings is balanced against reserving seats for potentially higher-paying individual passengers. This paper specifically addresses the need for a joint decision model that considers both group booking limits and dynamic pricing for individuals, a gap not fully explored in existing literature.
Methods
The authors employ a dynamic programming approach to model the joint decision-making process. They define a time-homogeneous model where orders arrive sequentially and can be either individual or group. For individual passengers, dynamic pricing is used to maximize revenue based on their willingness to pay. For group passengers, a fixed discount is applied, but the crucial innovation is the dynamic control of the booking limit for these group tickets. The model calculates the optimal price for individual tickets and the decision to accept or reject group bookings at each time step. An inverse recursive algorithm is developed to solve this dynamic programming model.
Datasets
Not discussed in the paper. The paper relies on simulated data for numerical experiments.
Key Insights and Ideas
- The "Symmetry Point" is Key: The paper emphasizes finding a balance between group discounts and individual pricing, framing it as finding a "symmetry point" for expected revenue.
- Dynamic Group Limits Outperform Fixed Limits: Dynamically adjusting the number of group tickets allowed is shown to be more effective than setting a fixed quota, allowing for greater revenue optimization.
- Group Discounts Shine in Weak Demand: The proposed joint policy is particularly beneficial when individual passenger demand is low, as group discounts can effectively stimulate demand and fill seats that might otherwise go empty.
- Individual Price is Tied to Group Decision: The optimal price for individual tickets is influenced by the decision to accept or reject group bookings, highlighting the interconnectedness of these strategies.
Key Results
Numerical experiments demonstrate that the proposed comprehensive model (CD) consistently outperforms a simple dynamic pricing model without group discounts (SD). The expected revenue of CD is higher than SD, especially when the overall arrival probability ($\alpha$) is low and the proportion of group orders ($\beta$) is high. The paper also shows that dynamically adjusting group booking limits (CD) leads to higher expected revenue compared to a dynamic pricing model with fixed group inventory (DF). The results suggest that as the individual demand-supply ratio increases, the advantage of the joint model diminishes, and SD can become slightly superior.
Limitations and Future Work
The authors acknowledge that the model assumes time-homogeneous arrival rates and that overbooking, no-shows, and refunds are not considered. Future work could explore more complex scenarios, such as multiple trains and stops, competitive factors, nonhomogeneous demand, and constrained information. Additionally, incorporating local or national conditions and analyzing the financial effects of specific policies (e.g., for unemployed or large families) are suggested avenues for future research.
Summary updated