Intelligent Genetic Algorithm with a Gradient-Based Local Search Applied to Supersonic Wing Planform Optimization

Abstract

AbstractThis study proposes an efficiently evolutionary algorithm, termed IGA-LS, which combines an intelligent genetic algorithm (IGA) with a gradient-based local search (LS). The IGA performs a crossover operation and uses a fractional factorial design to determine the optimal combination of design variables. That is, in the proposed IGA, the chromosomes of children are generated via an intelligent crossover process with factorial experiments after the execution of the selection operation of the GA. This gene mating process differs from that of the traditional GA, in which each chromosome of child exchange genes randomly. The gradient-based optimization approach seeks feasible and usable directions in which to minimize a specified objective function. The initial conditions are provided by the IGA in each generation. Therefore, the IGA-LS offers fast convergence and high numerical accuracy. Several multi-modal test functions are introduced herein to examine the algorithmic capacity and efficiency of finding the global optima. Moreover, the proposed IGA-LS algorithm is applied to optimize the design of a supersonic wing planform for supersonic airplane. The objective function is to minimize the drag during supersonic cruising. From the numerical results, the presented algorithm outperforms the traditional GA, the micro-GA and the intelligent GA in terms of convergence rate and value of the optimal function.