Artificial Intelligence through Evolutionary Programming: Prediction and Identification

Abstract
This report examines the manner by which evolutionary programming treats an arbitrary prediction problem. Additional experiments were conducted to clarify uncertainties given increased machine size and the cost/benefit of retaining offspring programs, the impact of noise on the predictive capability of the evolutionary process, and the efficacy of crossover as a mechanism for improving simulated evolution. It was also found that some difficult combinatorial problems such as the classic Traveling Salesman Problem can be addressed through less complex logics.

This publication has 0 references indexed in Scilit: