I. Introduction
Genetic algorithm (GA) is a powerful population based search procedure inspired by evolution theory. It proved to give good results when applied to different applications [1]–[6]. One of these applications is the feature/variable selection problem [2, 7–10]. In such a problem, a number of variables , that perform the best under certain classification scheme, are selected from a pool of variables . GA search consists of several steps: random population generation, fitness evaluation, fitness ranking, parent's selection, crossover and mutation operators, and a stopping criterion. A binary version of genetic population can be expressed as a list of binary strings with length in which the presence of a feature is expressed by ‘1’ and its absence is expressed by ‘0’. As an example, consider a binary string with six features ‘110101’, this represents selecting the first, second, fourth and sixth features.