Abstract | Keywords | Export | Availability | Bookmark

Choose an application

• Reference Manager
• EndNote
• RefWorks (Direct export to RefWorks)

The purpose of this work is twofold. In a first part, we developed a real-time strategic card game from scratch. It included the design of game rules, a user interface, cards, decks, and the implementation of the game in java with neither a game engine nor an external library. The game is playable online or offline, against other human players or artificial players, and with pre-built or custom decks.
In a second part, two artificial players were created for playing the card game. Firstly, a heuristic player that evaluates moves by quantifying the amount of resources it gains (or loses) versus the amount of resources its opponent gains (or loses). Secondly, a more advanced player that simulates the long-term consequences of a move until the end of the game to predict its outcome.
Results have shown the simulating player to outperform the heuristic one. By predicting the long-term consequences of moves, the simulating player has a stronger ability to planify its strategy, whereas the heuristic player always chooses the move maximising its short term reward. In addition, the simulating player’s performances are sufficient to challenge an expert human player.

Artificial intelligence --- AI --- Collectible card game --- CCG --- Ingénierie, informatique & technologie > Sciences informatiques