Where's Pikachu: Route Optimization in Location-Based Games

摘要

Along with the sudden rise in popularity of location-based games, the demand for tools to assist players in performance optimization increased. One major aspect lies within the analysis and planning of routes that contain a high amount of desirable game locations for the given player. However, personalized routes cannot be created by hand due to the amount of available in-game locations and the associated time constraints for real-world travel. This paper presents a system to dynamically create personalized route for players, based upon previous game data. These routes can be fully customized with regards to their time and location as well as the player's desired in-game goal, allowing them to e.g. specifically target their favourite species or maximize the amount of visited locations. The system is evaluated using a dataset of Berlin containing over 30.000 distinct locations with different associated in-game behaviour. Regarding the system's performance it is designed to work as an assistance system on a mobile device to assure its applicability in the context of location-based games.