A New Intelligent Genetic Algorithm for Aerodynamic Optimizations of Advanced Fighters

Abstract

In this paper, we proposed a new intelligent genetic algorithm by using a fractional factorial design to determine the best combination of design variables. In other words, in our genetic algorithm the children's chromosomes are generated by way of an intelligent crossover process with factorial experiments to select good genes from parents'. It is different from a tradi tional ge netic algorithm that the child's chromosome is usually generated by randomization. For a traditional genetic algorithm, such as simple genetic algorithm, coupling this kind of factorial design will enhance the performance with regard to evolutionary efficiency. In this paper, another type of intelligent genetic algorithm using Taguchi's orthogonal arrays is also developed for comparing with present algorithm. We present ed the re sults of five representative test cases including of three types of multi-modal function, a nonlinear dynamic control function, and a profile fitting for a high lift airfoil. All of the test cases are computed by using a traditional genetic algori thm, a micro genetic algorithm, the intelligent genetic algorithm using Taguchi's orthogonal arrays and our proposed genetic algorithm to evaluate the capability and efficiency of the four genetic algorithms. From the convergence histories and global optima, present genetic algorithm is showed superior to others. Further, present new intelligent genetic algorithm and other three algorithms are applied to the aerodynamic optimizations of two advanced fighters. Through the optimization analyses, we displayed the original and optimal wing planforms of the two fighters for comparison, and also depicted the drag force distributions, convergence histories to evaluate the evolutionary efficiency of the four genetic algorithms.