Artificial Intelligence for Game Developers

(10 Week Lesson Plan)

Lesson 1: Introduction to AI and Pathfinding

Textbook: pgs. 1 21 (Chapter 1)

Goals:

In this lesson, students are introduced to the initial concepts of AI and pathfinding. Well
begin with a few guidelines for AI in games, and delve into some quick descriptions of different
types of AI that exist. We will then move on to one of the fundamental types of AI: pathfinding.
We will explore the different types, and work through some specific examples of these types.

Key Topics:

General Principles of AI
The KISS Method
Hard != Fun
Play Fair

Types of AI
Pathfinding
Decision Making
Classification
Life Systems

Types of Pathfinding
Non-Look-Ahead Iterative Traversals
Look-Ahead Iterative Traversals
Look-Ahead Recursive Traversals

Non-Look-Ahead Iterative Traversals

Random Backstepping
Obstacle Tracing

Projects:

Lab Project 1.1: Pathfinding

Exams/Quizzes: NONE

Recommended Study Time (hours): 10 - 12

Lesson 2: Look-Ahead Iterative Methods of Pathfinding

Textbook: pgs. 21 55 (Chapter 1)

Goals:

In this lesson, students will continue their basic training in the core types of pathfinding: Look-
Ahead Iterative traversals. We will cover the common types of traversals including Breadth
First Search, Best First Search, and Dijkstras search. We will also discuss the Depth First
Search.

Key Topics:

Look-Ahead Iterative Traversals

Breadth First Search
Best First Search
Dijkstras Search

Look-Ahead Recursive Traversals

Depth First Search

Projects:

Lab Project 1.1: Pathfinding (cont.)

Exams/Quizzes: NONE

Recommended Study Time (hours): 10 - 12

Lesson 3: Data Structures

Textbook: pgs. 1 25 (Chapter 2)

Goals:

In this lesson, students are introduced to the famous A* search algorithm. We will talk about
how it works, its limitations, and how to make it more efficient. We will discuss heuristics, as
well as how we might apply A* to a simple RTS type game. Finally, we will discuss Hierarchical
pathfinding, and look at some examples of how it can be used.

Key Topics:

A* Search
How A* works.
Limitations of A*
Making A* more efficient
Heuristics

Hierarchical Pathfinding

Projects:

Lab Project 1.1: Pathfinding (cont.)

Exams/Quizzes: NONE

Recommended Study Time (hours): 8 - 10

Lesson 4: Non Gridded Pathfinding and The Pathfinding Demo

Textbook: pgs. 25 39 (Chapter 2)

Goals:

In this lesson, students will be exposed to the concept of pathfinding on Non-Gridded maps.
Methods for dealing with non-gridded worlds will be discussed as well. Finally, the student
will examine the design strategy used for the Pathfinding Demo, and learn how the interface
could be extended for their own use.

Key Topics:

Pathfinding on Non-Gridded Maps

Grid It
Visibility Points / Waypoint Networks
Radial Basis
Cost Fields
Quad-Trees
Mesh Navigation

Pathfinding Demo Design Strategy

Class Hierarchy
Interfaces
MFC Document/View Architecture

Projects:

Lab Project 1.1: Pathfinding (cont.)

Exams/Quizzes: NONE

Recommended Study Time (hours): 12 - 15

Lesson 5: Flocking

Textbook: pgs. 1 46 (Chapter 3)

Goals:

In this lesson, we will start to move from AI Pathfinding to AI Decision making. We will
discuss behavior based movement, and how to apply it in the case of the traditional
implementation of Flocking.

Key Topics:

Behavior Based Movement

Common Flocking Behaviors
o Separation
o Cohesion
o Alignment
o Avoidance
Other Possible Behaviors

The Flocking Demo

Design Strategies
Our Behaviors
o Separation
o Cohesion
o Alignment
o Avoidance
o Cruising
o Stay Within Sphere

Projects:

Lab Project 5.1: Flocking

Exams/Quizzes: NONE

Recommended Study Time (hours): 15 - 18

Lesson 6: Decision Making

Textbook: pgs. 1 24 (Chapter 4)

Goals:

In this lesson, students will continue their training in decision making started in the last
lesson, by learning about the different types of decision making commonly used in AI systems.
We will discuss one of the most popular types of decision making systems: state machines.
This will include an examination of transition diagrams, uses for state machines, and an
example implementation.

Key Topics:

Decision Making
Decision Trees
State Machines
Rule Base
Squad Behaviors

Finite State Machines

Transition Diagrams
Uses for State Machines
o Animation
o Game State
o Save File System
o A.I.

State Machine Demo

Design Strategy

Projects:

Lab Project 6.1: State Machines

Exams/Quizzes: NONE

Recommended Study Time (hours): 8 - 10

Lesson 7: Scripting

Textbook: pgs. 24 47 (Chapter 4)

Goals:

In this lesson, we will take a look at how to add scripting to our games. Well discuss a popular
scripting language called Python and learn we can integrate it into our state machine developed
in the last lesson.

Key Topics:

Scripting
Intro to Python
o Default Types and Built-ins
o Classes
o Functions
o Control Statements
o Importing Packages
Embedding Python
o Boost.Python
o Our Scripting Engine

Projects:

Lab Project 7.1: Scripted State Machines

Exams/Quizzes: NONE

Recommended Study Time (hours): 8 - 10

Lesson 8: Waypoint Networks

Textbook: pgs. 1 25 (Chapter 5)

Goals:

In this lesson, we will take a look at waypoint networks, which bring together both navigation
and decision making. We will look at how to create waypoint networks and how to load and
store them in our engine. We will also talk about the different kinds of data that can be stored
in waypoints and how we can use that information to help the AI make decisions. Finally we
will look at how to traverse the waypoint network using some of our earlier pathfinding
methods for the purposes of getting around in the environment.

Key Topics:

Waypoint Networks
Loading and Storing Waypoints
Edges
o Unidirectional vs. Bidirectional
o Cost Modifiers
Waypoints and Decision Making
o State Machine Updates
o Waypoint Orientation
Waypoints and Navigation
o A* Traversals

Projects:

Lab Project 8.1: Waypoint Networks

Exams/Quizzes: NONE

Recommended Study Time (hours): 8 - 10

Lesson 9: Putting it all Together

Textbook: NONE

Goals:

Before wrapping up the course, we will pull together all of the various topics we have covered
into a final application that makes use of everything. We will apply state machines to make
decisions, and pathfinding routines for traversing waypoint networks in the world. We will also
implement some simple squad level behavior and have a squad leader AI direct the entities
under its command.

Key Topics:

Final Demo Implementation

Pathfinding Algorithms
State Machines
Scripting
Flocking
Waypoint Networks
Squad Behavior Implementation

Projects:

Lab Project 9.1: Final Demo

Exams/Quizzes: NONE

Recommended Study Time (hours): 8 - 10

Lesson 10: Exam Preparation and Course Review

Textbook: NONE

Goals:

In this final lesson we will leave the student free to prepare for and take their final
examination. Multiple office hours will be held for student questions and answers.

Key Topics: NONE

Projects: NONE

Exams/Quizzes: NONE

Recommended Study Time (hours): 15 20

Final Examination

The final examination in this course will consist of 50 multiple-choice and true/false questions
pulled from all of the textbook chapters. Students are encouraged to use the lecture
presentation slides as a means for reviewing the key material prior to the examination. The
exam should take no more than two hours to complete. The final exam will be worth 100% of
student final grade.
Course Lab Projects

Chapter 1 2 Projects

Name: Pathfinding Demo

This application allows the user to make height fields using a grid interface, place costs on each
grid node, and visualize the path in both a 3D view as well as on the grid. The user is allowed
to choose the method of the pathfinding.

Application Type: Win32 Application (MFC)

Required Software: Visual Studio C++ Compiler (.net 2003), DirectX 9.0+ SDK
Required Hardware: x586 / 64MB RAM / 100MB Disk

Chapter 3 Projects

Name: Flocking Demo

This application allows the user to visualize fish swimming in schools, and dynamically adjust
the parameters of their Flocking algorithm.

Application Type: Win32 Application (MFC)

Required Software: Visual Studio C++ Compiler (.net 2003), DirectX 9.0+ SDK
Required Hardware: x586 / 64MB RAM / 100MB Disk

Chapter 4 Projects

Name: State Machine Demo

This application allows the user to design and visualize state machines and their execution. It
also provides for the user being able to dynamically create new scripted transition and action
states using Python.

Application Type: Win32 Application (MFC)

Required Software: Visual Studio C++ Compiler (.net 2003), ActiveState Python 2.3
Required Hardware: x586 / 64MB RAM / 100MB Disk
Chapter 5 Projects

Name: Waypoint Networks Demo

This application demonstrates the interaction between state machines, pathfinding, and
waypoint networks.

Application Type: Win32 Application (MFC)

Required Software: Visual Studio C++ Compiler (.net 2003), DirectX 9.0+ SDK
Required Hardware: x586 / 64MB RAM / 100MB Disk

Final Project

Name: Final Demo

This application has a squad leader entity directing other entities to go to various positions on
the map. The units traverse the map using a waypoint network.

Application Type: Win32 Application (MFC)

Required Software: Visual Studio C++ Compiler (.net 2003), DirectX 9.0+ SDK
Required Hardware: x586 / 64MB RAM / 100MB Disk