MPIA–Artificial Intelligence in Games
(January–Mars 2025)
Instructor:
Tiago de Lima
Email: tiago.delima@univ-artois.fr
Office: Faculty of Sciences C306
Description
This course is part of the module "Game Physics and Artificial Intelligence (MPIA)".
| Level: | 2nd year of the Masters degree |
| Credits: | 5 ECTS (for the complete module) |
| Lectures: | 10 h (10 × 60 min) |
| Labs: | 25 h (10 × 150 min) |
You will learn concepts, models and techniques for programming the behaviour of game elements. It focuses mainly on the artificial intelligence of games.
Topics to be covered:
- Game agents (non-playable characters):
- Perception, game viewed by agents.
- Mental states and social: beliefs, emotions, utilities, opinions, reputation, etc.
- Ad-hoc behaviour (finite state machines, behaviour trees)
- Markov decision processes
- Extensive games, stochastic games
AI Algorithms in games:
- Steering (chasing, evading, etc.), crowd effects (flocking, etc.)
- Search (A*, IDA*, hierarchical A*, etc.)
- Motion planning : configuration spaces (static vs. dynamic), spacial structures (navigation meshes, Voronoi diagrams, probabilistic road maps)
- Strategic decisions: goal-oriented planning, min-max decision, alpha-beta pruning, monte-carlo tree search, etc.
- Tactic decisions: “influence grids”, HMMs, etc.
- Collective decisions: utility aggregation, plan aggregation, cooperative games, etc.
- Planning (path planning, classical planning)
Prerequisites: Solid notions of algorithms and computer programming. The student is supposed to be able to write software in C++ and to use a game engine such as Unity or Unreal.
Schedule
Consult the timetable at ADE regularly. Course hours and rooms may change from one week to another.
| date | lecture | lab |
|---|---|---|
| 6 Jan | Vectors; Steering: | Steering library (submission: Feb 16) |
| 13 Jan | Ad-hoc behaviour: | |
| 20 Jan | Search: | |
| 27 Jan | Path finding: | |
| 3 Feb | Goal-oriented behavior: | |
| 10 Feb | Game theory: | |
| 24 Feb | Game theory (cont.): | Planning and path finding (submission: Mar 16) |
| 3 Mar | Game theory (cont.) | |
| 10 Mar | Learning | |
| 17 Mar | Project defense |
Evaluation
The assignments will be graded considering the correction and style of the code. Late submissions will be accepted with a penalty of 2 points per day. This means that an assignment submitted 10 days after the due date will be graded zero.
The evaluation of the course will be calculated based on the assignments as follows.
Physical engines assignments (M. Serf part):
AI in Games assignments (M. de Lima part):
Session 1:
Session 2:
About
The Moodle page of the course will be used for announces, lab project submissions and to communicate grades.
References
- Steering Behaviors For Autonomous Characters (by Craig W. Reynolds)
- A very good text about several different steering behaviours in games.
- AI for Games (by Ian Millington)
- Core Techniques and Algorithms in Game Programming (by Daniel Sánches-Crespo Dalmau New Riders)
- Chapters 6, 7 and 8.
- Artificial Intelligence and Games (by Georgis N. Yannakakis and Julian Togelius Springer)
- Chapter 2
- Artificial Intelligence: A Modern Approach (by by Stuart Russell and Peter Norvig)
- Simply the leading textbook in Artificial Intelligence and one of the most used textbooks in computer science worldwide.