Análise comparativa de heurísticas de tomada de decisão no jogo Exploradores

Abstract

This study provides a comparative overview of different heuristic approaches applied to the card game Exploradores (also known as Lost Cities worldwide). The main goal was to develop a virtual model of the game in C++, creating three types of virtual players with distinct methods for deciding their moves. The study involved simulating one thousand matches, in an automated manner, between combinations of players, allowing us to evaluate how well each strategy performed. The collected data included the execution time of each match and the total points scored by each player, enabling a detailed examination of the efficiency and effectiveness of the heuristics used. The results suggest that more direct methods, such as random choices, can outperform very simple planning attempts, while heuristics that consider more information about the current state of the game show superior overall performance in terms of points and consistency. This research contributes to understanding the impact of different strategies in the game, highlighting the value of using more elaborate heuristics to maximize scores.