HYBRID GENETIC ALGORITHMS AS TOOLS FOR COMPLEX OPTIMISATION PROBLEMS

摘要

In this work a hybrid optimisation procedure obtained by coupling a genetic algorithm and a gradient based optimisation routine is presented. After a brief introduction to genetic algorithms, the strategies which have been chosen for hybridisation are described. Results are then presented regarding both a single objective problem, consisting in the minimisation of a mathematical test function, and a multiobjective one, regarding an aerodynamic inverse design problem.