[4K3-GS-10] AI application: Movement

[4K3-GS-10-01] Prediction of Theme Park Waiting Times and Optimal Route Search using Gradient Boosting and Evolutionary Computation

〇Kazuki TAKEMI¹, Takuto SAKUMA¹, Shohei KATO¹ (1. Graduate School of Engineering, Nagoya Institute of Technology)

Keywords:Evolutionary Strategy, Gradient Boosting, Theme Park, Route Search, Wait Time Prediction

In this paper, we propose an application designed to predict wait times for attractions within Tokyo Disney Resort and provide routes that minimize both wait times and travel distances. Wait times and travel distances at these theme parks are generally long. However, by strategizing visitors' actions, it's possible to minimize these durations. Recently, the ability to check current wait times through official applications and share efficient touring strategies via social media has become available. Yet, the development of applications that construct efficient routes considering individual visitor preferences and dates has not progressed.
To address these issues, we propose the "TDL/TDS AI Navi". With the "TDL/TDS AI Navi", users simply select the attractions they wish to ride, and the system proposes an efficient route. Gradient boosting is used to predict wait times, and evolutionary strategies are employed to optimize the route based on the predicted wait times. As a practical verification, using the "TDL/TDS AI Navi" for a tour of 8 attractions within Tokyo Disney Sea resulted in a time reduction of 2 hours and 41 minutes.