Swarm algorithm with adaptive mutation for airfoil aerodynamic design

The Particle Swarm Optimization (PSO) method is sensitive to convergence at a sub-optimum solution for complex aerospace design problems. An Adaptive Mutation-Particle Swarm Optimization (AM-PSO) method is developed to address this challenge. A Gaussian-based operator is implemented to induce particle search diversity with probability through mutation. The extent of mutation during the optimization phase is governed by the collective search patterns of the swarm. Accordingly the proposed approach is shown to mitigate convergence at a sub-optimum design while concurrently limiting the computational resources required during the optimization cycle. The swarm algorithm developed is successfully validated on benchmark test functions with results favorably compared against several off-the-shelf methods. The AM-PSO is then used for airfoil re-design at flight envelopes encompassing low-to-high Mach numbers. The drag performances of the optimum airfoils are lower than the baseline shapes with the design effort requiring minimal computational resources relative to the optimization method documented in the literature.