Mr. Hollander discussed his undergraduate research in the areas of Genetic
Programming and Constrained Genetic Programming. Genetic Programming utilizes
the principals of Darwinian natural selection to breed populations of computer
programs. Initially, a random population of computer programs is produced. Once
the initial population of individuals is produced, Genetic Programming uses the
processes of Darwinian natural selection, sexual recombination, and mutation to
evolve populations of progressively fitter individuals. Fitness is defined as
the ability to find solutions to the problem at hand. Constrained Genetic
Programming allows the user to utilize problem specific knowledge or hypotheses
about the structure of potential solutions to specify constraints for the
creation of the genetic programs and thereby reduce the search space of the
problem. This can reduce the amount of resources spent searching redundant or
invalid spaces and thus can speed the evolution of more efficient solutions to
the problem.
