Future Works

An interesting problem might be to take Samuel’s checker playing program and incorporate the GA into the parameter searching routine. Samuel did a random search for his program. Even though it worked rather well, it took a lot of time for the search. By incorporating a GA, this search could be done more efficiently.

Also, the experience gained from this project could be used to attack a more complicated game. There are games that traditional AI methods do not play very well, such as bridge, poker, or go. These games represent difficult problems because of the enormous search space involved or because of the inexact information associated with the games. To come up with the right strings to represent and simplify those game strategies is the key to solving these problems.

REFERENCES

1  Bagley, J. D. (1967). The behavior of adaptive systems which employ genetic and correlation algorithms. Doctoral dissertation, Ann Arbor: The University of Michigan.

2  Ginsberg, M. (1993). Essentials of artificial intelligence. New York: McGraw Hill. pp. 49-99.

3  Goldberg, D. E. (1989). Genetic algorithms in search, optimization, and machine learning. Reading, MA: Addison-Wesley.

4  Haykin, S. (1994). Neural networks - A comprehensive foundation. pp. 397-434. New York: McGraw Hill.

5  Michie, D. (1962). Trial and error. New York: McGraw Hill.

6  Samuel, A. L. (1959). Some studies in machine learning using the game of checkers. New York: McGraw Hill.