ABSTRACT
In game theory, Minimax is an algorithm that aims to solve decision
making, based on a mathematical method that aims to minimize the
possible maximum loss, also known as maximizing the minimum
gain. The present work was developed as a research project and
investigation of algorithms in an itinerary of computational
thinking aimed at high school students. The research presents a
customized implementation of the algorithm in a “tic-tac-toe”,
where the Minimax algorithm is responsible for the AI moves that
oppose the player. The implementation was done in JavaScript
(WEB) and tests were performed with different equations in the
definition of the nodes' score. With the tests, it was possible to
perceive which steps and situations most impact the quality of the
moves, as well as validate different parameters for the algorithm
depending on the desired degree of difficulty. In the end, tests were
also carried out in matches against real players, and it was possible
to perceive that Minimax can provide different strategies and levels
of difficulty in the tic-tac-toe game.