Sie sind auf Seite 1von 2

Artificial Intelligence for Game Design

Lernziel

The goal of this course is to teach Artificial Intelligence (AI) techniques in computer
game design and development for computer scientists.

In this course the student will learn the principles of AI and how can they be applied in
the field of computer game design, the most popular techniques and technologies for
game implementation, and the many ways in which advances in computer graphics and
artificial intelligence influence game design.

This two days course is based on two AI Game Engines Asteroid and Raven; both of
them are developed in Visual C++ and have all components of modern Game Engines
such as path finding and different techniques of decision making. They already have
animation, graphics in 2D and 3D, sound etc. However in this part we are not concerned
with these last aspects. We are interested only in the AI-game Engines techniques. This
course is composed of 8 chapters and will consist of lectures and laboratory work.

Kursinhalt

Chapter 1: Introduction to AI Game Engine


In this chapter, students will be familiarized with AI concepts applied to games. They will
learn basic components and design of AI-Game Engine. We begin with a description of
decision making techniques used in game design as well as pathfinding.

In the first lab session, students will experiment with AI-game Engine called Asteroids.
In the second half of the lab will make first contact with pathfinding applications or
behavior based movement such as: Seek, Flee, Arrive, Pursuit and Evade.

Chapter 2: Pathfinding techniques


In this chapter, students start discovering the working principles of pathfinding such
Breadth First Search, Depth First Search, Best First Search, Dijkstra's, Heuristics and A*
search.

In the lab session, students will experiment and implement the following apllications:
Wander, Pobstacle Avoidance, Interpose, Hide, and Wall Avoidance.

The source code of the following pathfinding searches: Breadth First Search, Depth First
Search, Best First Search, Dijkstra's and A* search will be distribued ,commented on,
discussed, modified, compiled and tested. The students will learn how pathfinding can be
used.

Chapter 3: Autonomously Moving game agents, Squads


This chapter is concerned with two aspects:
a. Working principles of Autonomously Moving game agents
b. Common Squads behaviours: How can many agents move into the same direction,
accomplish together a task or collaborate to work in order to solve a problem.

In the lab session students will experiment squads and will put the techniques learned in
chapter 1 and chapter 2 to work together. This task is realized with the famous game
called Raven.

Again, students will get the working source code of the Raven2, experiment it, modify
the code, implement their own idea and test it.

Prof. Dr. Taoufik Nouri, Fachhochschule Nordwestschweiz, Hochschule für Technik Seite 1/2
Chapter 4: Decision Making: Decision Tree, State Machine, Rule based
In this chapter, students start learning different kind of decision making commonly used
in AI-Game design such as State Machines with their transition diagrams, decision Trees
and Rule Based.

During the lab session, students will examine and implement State machine, decision
trees and Rule Based System. Students will master the use of these techniques in game
design.

Chapter 5: Decision Making: Fuzzy logic FUSM


In this chapter will experiment with Fuzzy-state machines (FuSMs) components. FuSMs
are built on the notion of fuzzy logic. Fuzzy Logic is defined as a superset of conventional
(Boolean) logic that has been extended to handle the concept of vagueness. In game
design as well as in live things are not true or false. Things are not dual. There is a
common region between true and false.

In the lab session, students will implement the notion of FuSms in the Raven2 Game and
compare it to the traditional Game Engine used in part 3.

Chapter 6: Game Design: Soccer-Project


In the lab session students will experiment/implement/change/adapt/use distributed
techniques with the projects Soccer.

Chapter 7: Advanced AI Game-Engine Techniques: Genetic Algorithms


Genetic algorithms are starting to show up in games, fresh from the academic world.
They are used especially in cases where a traditional algorithm is not any more useable.
Students will learn the principles and the use cases of Genetic Algorithms.

In the lab session students will experiment two implementations of Genetic Algorithms:
one in the Raven game Engine and the second in the Asteroid Game Engine.

Chapter 8: Advanced AI Game-Engine Techniques: Neural Networks NN


Student will get an introduction to the working principles as well as the different types of
NN and their use cases.

In the lab session students will experiment many types of NN implemented in the Raven2
game Engine and the Asteroid Game Engine.

Prof. Dr. Taoufik Nouri, Fachhochschule Nordwestschweiz, Hochschule für Technik Seite 2/2

Das könnte Ihnen auch gefallen