Surface shape design in fluid flow problems via hybrid optimization algorithms

Shape design problems, in general, and inverse design problems, in particular, are often solved via optimization techniques. Evolutionary algorithms provide robust and efficient solution methods for such problems. This paper focuses on the application of genetic algorithms (GA), particle swarm optimization (PSO), and two hybrid variants of GA and PSO. Optimum shapes in five shape design problems are found by the proposed hybrid algorithms. Potential, Euler and both laminar and turbulent Navier–Stokes flow solvers are employed in the test problems which include internal and external flows and convection heat transfer. Computational results show that hybridization of GA and PSO improves the convergence rate in all test cases. Up to 30% speed up is observed in the numerical test cases when the hybrid methods are employed and it is also shown that hybrid methods find a better solution in the design space as compared to either GA or PSO.