Intelligent Agent Implementation: Comparison of Search Strategies in Pokémon Unite.

Abstract

In the context of a practical assignment for the "Artificial Intelligence" course, an intelligent agent was developed capable of playing a simplified version of "Pokémon Unite", where the agent navigates a map to defeat opposing pokémons, reach the "master Pokémon," and defeat it to win. The FAIA framework was used to implement the agent, and a graphical user interface was created
using FXGL and JavaFX to visualize the game. This paper provides a detailed description of the solution, including the representation of the game scenario, the definition of the agent's state and goal, how the agent obtains its perceptions, and the actions that it can perform. Finally, the agent's performance was evaluated using different search strategies such as BFS and Greedy in various controlled scenarios, recording the agent's final energy after achieving the objective.