This thesis is an exploratory study on airline ticket purchasing. Actual data are collected and studied with the goal of using a technique to purchase the lowest cost airline ticket given the information available. These data are studied in two ways: an algorithm based on the solution to the Secretary is applied and the trends of the data are reviewed. The Secretary Problem is where an employer is interviewing candidates for a secretarial position and the response to hire or reject a candidate must be made immediately after the interview. A sensible strategy to select the best candidate would be to interview and reject candidates without consideration to get an idea of the pool, and then at some point begin searching for the best candidate based on the initial pool. Generally, the first candidate found better than the initial pool is accepted. Buying an airline ticket can be in a way thought of as analogous to this: at each candidate ticket, one must deicide to accept and purchase the ticket or reject the current ticket price and wait for a potentially better ticket price. From the data, one can see that sudden drops in price make airfare purchasing a good candidate for the Secretary Problem.
The application of a modified solution of the Secretary Problem to include heuristics noted by industry experts is successful at saving money over simply buying the first ticket offered to a given customer in about half of the collected routes, as the other cases exhibit a continuously increasing behavior. In the cases it is successful, the ticket is purchased about 1.5 to 2 weeks before departure, during a price valley. In summary, the modified Secretary Algorithm is successful as it saves money for some of the customers.