Programming Abyss with AI

The objective of my Senior Project is to create a simplified version of a board game Abyss designed by Bruno Cathala and Charles Chevallier. By simplified I mean consisting only of one part of the original game, which is the Exploration phase. For the purpose of constructing the game I use the idea of a finite state machine implemented in Processing. Aside from the game itself, I have also created three different artificial players. Each of them makes decisions based on different types of calculations. Deterministic player takes into account part of a state in which the environment is in to determine its decision. Probabilistic player operates similarly to Dungeons and Dragons mechanics - it calculates probability of undertaking a specific action based on the state of the environment, and then picks a random number - if the number is lower than the probability, then the action is chosen. Utility player on the other hand, compares utilities of possible actions before making its decision. In this paper I analyze and compare 100 gameplays of these artificial agents to determine which type of artificial intelligence is the best at winning the gam