United States Department of Transportation
i
Passenger Demand Model for Railway Revenue Management
-
2011-01-01
[PDF-458.06 KB]
Details:
-
Creators:
-
Corporate Creators:
-
Corporate Contributors:
-
Subject/TRT Terms:
-
Publication/ Report Number:
-
Resource Type:
-
Geographical Coverage:
-
Corporate Publisher:
-
Abstract:In this paper, we have illustrated a fare pricing strategy for the Acela Express service operated by Amtrak. The RM method proposed is based on passenger’s preference and products’ attributes. Using sales data, a MNL model has been calibrated; the random utility theory has been applied to explain passengers’ choice of booking time under a range of hypothetical sale horizons. In order to capture aggregate passengers’ response to fare price, a demand function based on OLS regression has been incorporated in the procedure. This approach is appealing because it allows product attributes such as departure day of week, fare price and destination specific effects to be taken into account in the RM problem. The two models are incorporated in a mathematical formulation that maximizes the expected revenues for each departure day and for each destination market. Our analysis provides a method for estimating choice behavior and passenger demand in response to RM strategies from readily available booking data. The accuracy of the estimates depends on the market size; for instance, the model produces good results for station5 market which is the predominant market for Acela Express. Overall, we show that the proposed model in this paper is promising and can potentially lead to increase in revenue. It was demonstrated that the pricing strategy which accounts for choice behavior could potentially increase the revenue from 2.06 to 14.64 percent and 0.70 to 11.60 percent per day within the respective weeks of March and April. However, it should be noted that, as with any academic work, the model is based on some simplifying assumptions which might not fully comply with the real world problem. For example, Amtrak pricing strategy is more complicated than what presented in this paper. We did not account for cancellation behavior, various discounts, guest reward program, special fare plans or competition with non-Acela trains or other modes of transportation. Also, the choice model is not tested independently to show if it accurately reflects customers’ choice behavior in the market. So there is significant room to improve or extend this research. Several research extensions are suggested. The new pricing strategy should be tested in terms of market acceptance and pricing response. Due to lack of socioeconomic information from our sales data, it would be desirable to calibrate a latent class model by identifying different passenger segments in terms of trip purpose or socioeconomic characteristics. The model calibrated handles deterministic heterogeneity only. Mixed logit models could be adopted to address random heterogeneity in customer behavior. Both latent classes and random parameters logit models have the potential to improve the accuracy of the customer choice model. To conclude, our booking data can be used to study cancellation behavior for high quality rail services. The optimization routine based on choice behavior and different time horizons could be adopted by other operators that sell products on-line (i.e. shippers, couriers).
-
Format:
-
Funding:
-
Collection(s):
-
Main Document Checksum:
-
Download URL:
-
File Type:
