Genetic algorithm coupled with a deterministic method for optimization in electromagnetics

Abstract

International audienceIn this paper, a hybrid technique for global optimization based on the genetic algorithm and a deterministic method is presented. A potential advantage of the hybrid method compared to the genetic algorithm is that global optimization can be performed more efficiently. An intrinsic problem of the hybrid techniques is related to the moment of stopping the stochastic routine to launch the deterministic one. This is investigated using some natural criteria for the commutation between the two methods. The results show that it is possible to gain in efficiency and in accuracy but the criterion is usually problem dependent. Finally, to show the solution of a real problem, the hybrid algorithm is coupled to a 2D code based on the boundary element method to optimize a connector of 145 kV GI

Fundamentos, Potencialidades e Aplicações de Algoritmos Evolutivos

“…Os métodos estocásticos, dos quais os AEs fazem parte, não necessitam do cálculo do gradiente e são aptos a encontrar a solução global. Contudo, o número de avaliações da função objetivo, necessárias para encontrar a solução,é normalmente maior que o número requerido pelos métodos determinísticos [164].…”

“…Segundo Vasconcelos et al [164], a maior dificuldade no uso de metodologias híbridas, também conhecidos por algoritmos meméticos [112], [108] (apesar das controvérias em adotar este nome para este tipo de método) ou evolução Lamarckiana [169], [85], [118] …”

“…Assim,é impossível garantir que a soluçãoé a global. Vasconcelos et al [164] propõem os seguintes critérios para comutação entre AE e métodos de busca local que podem ser definidos pelo(a):…”

Gradient based optimization of Permanent Magnet generator design for a tidal turbine

“…GA have been widely used in optimization of electrical machine design [3], [6], [7], however, other evolutionary algorithms, like Immune Algorithms, Evolution Strategy, Differential Evolution [8], [9] and Particle Swarm Optimization [10] also show good results. In many cases, a hybrid solution is employed, where an Evolutionary Algorithm is used to find the area of the global minimum, and a local search algorithm is used to find the precise location of the minimum [4], [11], [12]. This reduces the number of evaluations drastically.…”

“…For many years, much focus has been put on optimization of electrical machines and on PM machines in particular [2]- [4]. [5] presents an overview of different optimization methods used for design of electrical machines.…”