A New Metaheuristic Method with Applications to Airfoil Shape Optimization

This paper proposes an efficient meta-heuristic method called expert groups’ optimization algorithm. The method strategy relies on four principles and starts from a random initial population. The population members are divided into two expert groups: the free group and the guided group. Each group has specific tasks for effective domain search, but with one new operator. This operator has an intelligent mechanism so that exploration and exploitation of the population can lead the members to the global optimum. The new method is validated through a standard test function. Then its performance is evaluated in the application of an inverse geometric reconstruction and the results are compared with a genetic algorithm, particle swarm optimization, and mean-variance mapping optimization. Results show that the new method outperforms the alternative methods in convergence rate and reaching the global optimum. Finally, the expert groups’ optimization algorithm performance is evaluated in an engineering problem with high computational cost. In this case, the goal is drag coefficient minimization of the RAE 2822 airfoil in transonic flow at a fixed lift coefficient with constraints on the pitching moment and airfoil area. An unstructured grid Navier-Stokes flow solver with a two-equation turbulence model is used to evaluate the aerodynamic objective function. The results show that the optimal solutions obtained by the new method outperform those of mean-variance mapping optimization with considerably faster convergence.