A typical GA solves unconstrained optimization problems, so that a traditional GA method was presented in which penalty functions have been utilized as the selection criterion of surviving individuals to apply GA for constrained problems. Sensitivity of the convergence properties of this method to the penalty parameter makes the try for a recently new method on title of “Modified Genetic Algorithm (MGA)”, with a special selection criterion to overcome this sensitivity difficulty; but this new method is not general to apply in any optimization problem; thus, in this paper, a method has been proposed to generalize the recent modified GA. Furthermore, in this method, the cost of iteration has been minimized by selecting a suitable termination check. In this paper, it also has been tried to compare the mentioned generalized modified GA with the particle swarm optimization (PSO) method. The last section of the article consists of application of this method in some famous optimization problems and a MDO (multidisciplinary design optimization) in aerospace field.

Keywords