 |
Dr. Robert G. Reynolds Biography
Wayne State University/University of Michigan-Ann Arbor
USA
"Using Cultural Algorithms to Design Controllers for Robot and Softbots"
A. Tutorial description:
Cultural Algorithms were developed by Reynolds as a computational framework in which to embed social learning in an evolutionary context [1979]. Unlike traditional learning approaches Cultural Algorithms derive their power from large collections of interacting agents. Within virtual worlds such as games or other interactive digital entertainment systems it is often the case that we wish to coordinate the behavior of large groups of intelligent agents in an efficient fashion. This tutorial focuses on the ability of Cultural Algorithms to perform large-scale group learning within these virtual worlds. They have been used to generate socially intelligent controllers and group social behavior in various gaming genres, both serious and fun. This tutorial describes Cultural Algorithms and how they can be used to incorporate social intelligence into a virtual world using examples form a variety of genres. A toolkit based upon the Cultural Algorithms paradigm will be presented and used as a basis for developing example applications.
B. Content Outline:
- What are Cultural Algorithms:
A basic description of the Cultural Algorithm Framework and its relationship to other socially motivated learning technologies will be described along with an introduction to the Cultural Algorithms Toolkit, CAT 3.1 that supports mutli-objective problem solving.
- Why Cultural Algorithms work:
Here we discuss the basic phases of the problem solving process in Cultural Algorithms and how those phases emerge from the interaction of the knowledge sources in the belief space, knowledge swarms, and the population of problem solvers in the population space. Convergence properties will be discussed.
- When will Cultural Algorithms work?
Since Cultural Algorithms derive their power from the emergence of knowledge and population swarms, what problems are suitable for solution with Cultural Algorithms and what problems will be hard or deceptive?
- Virtual Worlds:
Since Cultural Algorithms derive their power from the emergence of knowledge and population swarms, what problems are suitable for solution with Cultural Algorithms and what problems will be hard or deceptive?
a. The Basic Components
b. Opportunities for Socially Motivated Learning in Virtual Worlds.
1. Avatar design.
2. Collective movement.
3. Group decision making.
4. Level design.
5. Application Examples:
a. Game Worlds
1. Platform Games: Super Mario
2. 3D Racing Games
3. Real Time Strategy Games: Starcraft
b. Real Worlds
1. Emergence of Urban Centers
2. The Ancient Land Bridge (Prehistoric Hunter-Gatherers).
3. Vanishing Societies
4. Health Care Applications
5. Sustainable Culture
|
