Simulating Airline Behavior: Application for the Australian Domestic Market

Abstract

This paper demonstrates the ability of a model, which simulates competition between airlines in a domestic aviation market, to accurately reproduce real-world behavior. The Australian market was chosen as a test case as it is a geographically isolated region with significant demand and complexity, including one of the busiest routes in the world, where connecting international passengers do not significantly skew the market. The model is based on an n-player noncooperative game, in which each airline represents a player within the game. The primary assumption is that each airline attempts to maximize profits by adjusting the decision variables of airfares, flight frequency, and choice of aircraft on routes within its network. The approach works iteratively, allowing each airline to respond to the decisions made by other airlines during each successive optimization. The model is said to reach convergence when there is no significant change in any airline’s profit from one iteration to the next. Once this occurs, the predictions of each airline’s decision variables can be compared with real data. The model gives highly detailed predictions of airline specific airfares, flight frequencies on segments, passenger flows, and airline market share, which strongly correlate with observed values.