Genetic Algorithms (GA) have been widely used in the areas of searching, function optimization, and machine learning. In many of these applications, the effect of noise is a critical factor in the performance of the genetic algorithms. While it hals been shown in previous siiudies that genetic algorithms are still able to perform effectively in the presence of noise, the problem of locating the global optimal solution at the end of the search has never been effectively addressed. Furthermore, the best solution obtained by GA often does not coincicle with the optimal solution for the problem when noise is present. In this report, we describe a modified GA for dealing with noisy environments. We use an optimal solution list to keep a dynamic record of the optimal solutions that have been found during the course of evolutia~no f the population of noisy solutions. In addition, we also vary the population size and sampling rate to achieve further improvements. We demonstrate the performance of our scheme via a simple function optimization problem using genetic algorithm in a noisy environment. Our results show that the optimal solution list is able to provide a small solution set that provides near optimal solutions obtainable in the absence of noise. Our scheme is also easily implemented in practice with the addition of a simple optimal solution list and minor changes to the selection and evaluation phases of an existing GA implementation.
Date of this Version