Optimizing Fun: A TSP-Based Approach to Route Optimization at Disneyland

Abstract

Despite Disneyland sitting on a one square mile plot of land, the average distance walked by a guest per day is over seven miles. Many of these unnecessary miles are because the guests have no way to optimize their route based on their ride preferences and timing. Disneyland does offer solutions to help guests skip-the-line, but none to help guests optimize their day by minimizing walk time and prioritizing rides. This thesis applied the Traveling Salesman Problem to routes through Disneyland in order to minimize the walking distance of guests. Through a series of complications, we expanded the problem to better model a realistic day. First, dining breaks were included in the route. Then, guests were able to prioritize which set of attractions they are most interested in experiencing. Finally, downtime was included in the itinerary to allow for breaks and unavoidable delays. These three complications combine into a final route and itinerary for a day at Disneyland that decreases the walking distance by 70% while improving the guest experience. Because the resulting model adds significant value to a guest’s day, we analyzed the potential incremental revenue a personalized model could generate for Disneyland were it to be commercialized. After careful estimation of its potential value, we found that the commercialized model could increase revenue by up to 12% or $540 million per year.