 |
    |
Automatic Creation of Team-Control Plans Using an Assignment Branch in Genetic Programming
Walter A. Talbott
Stanford Symbolic Systems Program, Stanford University, Stanford, California 94305wtalbott@stanford.edu
Abstract. This paper is concerned with the introduction of a method for allowing genetic programming to automatically create team-control plans using an assignment branch. Team-control plans are representations of the composition and behavior of groups of agents and include the definition of one or more roles. Genetic programming is a general problem solving technique, and using genetic programming as a means for creating team plans would allow the user to specify very little, while producing effective results. I propose to show that the use of what I call an assignment branch in genetic programming provides a robust and general way to approach team composition and role definition problems.
LNCS 3102, p. 201 ff.
lncs@springer.de
© Springer-Verlag Berlin Heidelberg 2004