Organisation

Time only SS
Time Friday, 10.00h-13.15h
Room 056
Credits 4SWS/6ECTS
Exam written, 90min

Announcements

  • First lesson in SS 19: 22.03.2019

Contents

This lecture consists of 2 parts: A theoretical part, which introduces the AI-algorithm applied in game AI, and a practical part, where the application and implementaion of AI in modern computer games is demonstrated.

  • Pathfinding:

    • Algorithms for pathfinding: A*, Hierarchical Pathfinding, Jump-Point-Search, Rectangular Symmetry Reduction,
    • Monte Carlo Tree Search
    • Heuristics for Indoor- and Outdoor-Pathfinding
    • World Representations: Tile-Graphs, Nav-Meshes, Dirichlet-Domains

  • Decision Making and Planning:

    • Decision Trees
    • State Machines
    • Behavior Trees
    • Fuzzy Logic
    • Goal Oriented Behavior

  • Solving Constrained Satisfaction Problems (CSP)

  • Tactical and Strategic AI:

    • Waypoint Analysis
    • Tactical Analysis
    • Tactical Pathfinding
    • Coordinated Action

  • Machine Learning in Computer Games:

    • Categories and Applications in General
    • Conventional Approaches for supervised Learning

  • Reinforcement Learning:

    • Definitions, Categorization, Examples
    • Markov Decision Process
    • Modellbased Learning
    • Q-Learning
    • Exploitation/Exploration

  • Deep Reinforcement Learning:

    • Q-Network
    • AlphaGo
    • AlphaZero
    • AlphaStar

  • Procedural Content Generation:

    • Genetic Algorithm
    • Generative Adversarial Networks (GAN)

Structure, Contents, Documents

Gitlab Repo for this lecture

Lecture Contents Addtional Material
Introduction AI in general, AI in Games
Pathfinding Pathfinding algorithms for indoor- and outdoor environments, world representations
Decision Making I Decision Trees, State Machines, Behavior Trees
Decision Making II Fuzzy Logic, Goal Oriented Behavior, Goal Oriented Action Planning
Constraint Satisfaction Problems Constraint Satisfaction Problems, Backtracking, Forward Checking, Constraint Propagation A*-Example 8-Puzzle [.html] , A*-Example 8-Puzzle [.ipynb] , CSP-Example Sudoku [.html] , CSP-Example Sudoku [.ipynb]
Tactical and Strategic AI Waypoint Tactics, Tactical Analysis, Tactical Pathfinding
Monte Carlo Tree Search
Reinforcement Learning Value-Iteration, Q-Learning