Dr. James Glenn, Amherst College
Title: Algorithms and Approximations for Games of Strategy and Chance
Host: Holly Rushmeier
Psychologists, economists, and biologists are interested in games as models of behavior of individuals, organizations, and organisms; the games of interest to them often feature simultaneous movement. Researchers in artificial intelligence have long used combinatorial games such as chess, checkers, and go as testbeds for their techniques and inspirations for new approaches. We discuss a class of games with an element of strategy and chance like Yahtzee and Can’t Stop, algorithms for computing optimal strategies for them, applications of genetic algorithms to finding approximately optimal solutions, and strategies for dealing with the games’ inherent randomness when implementing the genetic algorithms.