Repository URL https://github.com/mimirerelala/ElRumDrinkingCapitan/edit/master/README.md
This is our implementation of an AI for Santase card game (66). The AI can be tested at http://ai.bgcoder.com
The source code and logic are forked from https://github.com/NikolayIT/SantaseGameEngine
Our code contribution can be found in: the directory ElRumDrinkingCapitan/SantaseGameEngine-master/Source/AI/mcts/
The logic is NOT complete, and does not yet perform as expected. We tried to implement a Monte Carlo tree search method, which implies searching through all possible game steps, simulating the game to the end, and then choosing the most profitable step and in the same way reevaluating the situation. Monte Carlo methods are a common tool in game AI, performing well in situations withouth perfect information.
For simulating the game itself we use a knowledge based approach, taking basic assumptions and simplifications of the game.
This project was part of a Data Structure and Algorithms Teamwork project in the 2015 edition of Telerik Academy - free software engineering trainig courses. More about the courses can be found at https://telerikacademy.com