Abstract - To study, model and enumerate intelligence, one must solve or approximate solutions to hard problems. Many methods have been advanced in this pursuit including simulated annealing , genetic algorithms , and evolutionary algorithms . The Bayesian network method studied and reported here is another such method for optimizing a function, which we call utility. We use the Bayesian network method asit is analogous to the way people perceive, categorize, and calculate on their environment . The Bayesian network method is a reasonable approximation to human reasoning process . The Decision-theoretic approach is used specifically to solve the `Herding Problem' but the method and software is extensible to any other hard, discreet, optimization problem. We have successfully created a simulation environment in which we will be able to explore cooperation among multiple, intelligent, learning-capable agents.
Date of creation, presentation, or exhibit
Department, Program, or Center
Microelectronic Engineering (KGCOE)
Tillett, Jason and Sahin, Ferat, "Intelli-agent: Simulation environment for cooperating intelligent agents" (2002). Accessed from
RIT – Main Campus