AI is My Co-Pilot: Did ChatGPT Just Solve the Oldest Travel Riddle?

Jeffrey Berthiaume
8 min readJan 29, 2024
Apple Vision Pro simulation showing mock-up of flight path monitoring

In an age where technology seamlessly intertwines with our daily lives, the realm of airline network management is no exception. In a moment of whimsy, I enlisted ChatGPT to address a classic challenge in optimization: the Traveling Salesman Problem (TSP).

At its core, the TSP involves determining the most efficient route that visits a series of locations and returns to the starting point. It’s a problem that has perplexed mathematicians and computer scientists for decades, yet it finds practical application in modern airline route optimization, especially in point-to-point networks.

Airline Network Models: Point-to-Point vs. Hub-and-Spoke

In the airline industry, choosing a network model is a pivotal decision. While the hub-and-spoke model centralizes traffic, the point-to-point model offers a direct route option, crucial for smaller airlines seeking operational efficiency and customer satisfaction. I chose to model this when generating some mock (or faked) traffic.

The Challenge

Crafting a realistic mock airline network was my primary goal. I needed data: flight schedules, aircraft types, city pairs — the works — over a series of years.



Jeffrey Berthiaume

Jeffrey Berthiaume is a technology veteran and senior executive, who has spent decades creating and innovating technology.