Fare Availability Inference Using Live Price Matching
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for determining airfare class availability are slow, energy-intensive, and require substantial data storage due to the need for frequent remote server access and storage of all possible queries, which is impractical for a large number of routes and airports.
Innovation Solution
A method and server that infer airfare class availability by analyzing observable live bookable prices with stored data, using statistical models and machine learning to predict fare classes without requiring frequent server access or extensive data storage, utilizing cached prices and historical patterns to estimate availability.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If availability queries are run against GDS systems to obtain accurate fare class availability, then measurement precision is improved, but loss of time increases and use of energy increases
Solution Approach 1:
The system pre-calculates and stores expected fare prices for all possible routes and dates using historical data and pricing models before queries are executed. This preliminary preparation allows the system to quickly compare actual GDS prices against pre-computed expectations without performing full availability calculations at query time, thereby reducing query execution time while maintaining accuracy.
Solution Approach 2:
The system introduces an intermediary pricing model that acts as a mediator between the GDS system and the user interface. Instead of directly querying GDS for every possible fare class availability, the system uses the pricing model to infer availability by comparing observed prices against expected prices, reducing the need for time-consuming direct GDS queries.
2Adaptability or versatility
If availability queries are run against multiple airline servers to obtain comprehensive fare information, then adaptability is improved, but use of energy increases and loss of time increases
Solution Approach 1:
The system merges multiple data sources including historical pricing data, current GDS prices, and pricing model predictions into a unified inference mechanism. By combining these sources and using statistical analysis to infer fare class availability, the system achieves comprehensive coverage without needing to query every airline server directly for every query, thereby reducing energy consumption.
3Adaptability or versatility
If fare class availability data is stored for all possible routes and time periods in GDS, then adaptability is improved, but quantity of substance increases
Solution Approach 1:
The system extracts only the essential pricing components and historical patterns needed for inference, storing compact representations of pricing models and historical data rather than complete fare class availability data for all possible routes. This extraction approach maintains adaptability by preserving the underlying pricing logic while significantly reducing storage requirements.
Solution Approach 2:
The system changes the representation parameters from storing complete fare class availability matrices to storing compact pricing models with key parameters such as historical price distributions, route characteristics, and pricing rules. This parameter transformation enables the system to generate availability information on-demand with minimal storage requirements while maintaining comprehensive route coverage.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
The field of the invention relates to methods, servers and computer program products for inferring which fare classes are available. There is provided a method of inferring fare class availability for a specific flight or flights by: (a) receiving observable, live bookable prices for that flight(s), in which that price information is not accompanied by complete fare class information for those flight(s); (b) comparing the observable, live bookable prices with a set of prices calculated from stored data, such as valid itinerary data, fare data such as FROP data, and tax/ surcharge data; and (c) determining which fare class was used in the observable live bookable price by determining which fare class was used in the calculated price that matches most closely to the observable live bookable price.