As in previous semesters, the goal is intelligent or quasi-intelligent behaviour in a simulated environment, a real-time computer-game. Although apparently "just a game", it turns out that much of the challenges apparent in REAL-WORLD applications of artificial intelligence are exhibited in this type of complicated simulated environments. Additionally, agent intelligence in computer games is also increasingly an application on its own right, as the multi-billion dollar computer gaming industry companies continue to attempt better simulated opponents for the newer computer games (Starcraft, DOOM-type games, etc.).
Students participating in the course will be working in groups of up to 3 people, where each group implements an integral part of the project. The project for this semester is simulating a variant of a game of HUNT, and designing (and implementing) agents that can exhibit intelligent game-play in that environment.
Class hours are intended for initial material and problem domain presentation, as well as design and design review (first 3 weeks), mid-project progress and tests (2 weeks in the middle of the semester), and final integration and delivery (last 2 weeks of the semester). Other weeks will have no official class hours, but are allocated for close supervision of the course instructor, of the programs under developement.
Note that since, this year, an AI course is NOT a prerequisite, it is still possible to do a non-AI project in this course (programming the environment simulator), or even a player, but using a "QUASI-AI" approach (to be explained in class).
Back to BGU - CS HomePage