Cheol-Gyun Lee

Multimodal function optimization based on particle swarm optimization

In this paper, a new algorithm for the multimodal function optimization is proposed, based on the particle swarm optimization (PSO). A new method, named the multigrouped particle swarm optimization (MGPSO), keeps basic concepts of the PSO, and, thus, shows a more straightforward convergence compared to conventional hybrid type approaches. Moreover, the MGPSO has a unique advantage in that one can search N superior peaks of a multimodal function when the number of groups is N. The usefulness of the proposed algorithm was verified by the application to various case studies, including a practical electromagnetic optimization problem

Niching genetic algorithm with restricted competition selection for multimodal function optimization

The niching method enables the genetic algorithm to be applied to the problems that require the location of multiple solutions in the search space. In this paper, a new niching method using restricted competition selection (RCS) is proposed to identify and search multiple niches (peaks) efficiently in a multimodal domain. To verify its validity, the proposed method is applied to some traditional mathematical problems and an induction motor design.

Induction motor design for electric vehicle using a niching genetic algorithm

In the case of the shape or structural optimization of induction motor design, it is necessary to identify multiple optimal profiles by locating local optima as well as global. Niching methods extend genetic algorithms to domains that require the location and maintenance of multiple solutions. In this paper, optimal design of an induction motor for an electric vehicle using a niching method adopting restricted competition selection is proposed. The evaluation criteria and the standards of the best design selection are also presented.

An Improved Particle Swarm Optimization Algorithm Mimicking Territorial Dispute Between Groups for Multimodal Function Optimization Problems

In the present paper, an improved particle swarm optimization (PSO) algorithm for multimodal function optimization is proposed. The new algorithm, named auto-tuning multigrouped PSO (AT-MGPSO) algorithm mimics natural phenomena in ecosystem such as territorial dispute between different group members and immigration of weak groups, resulting in automatic determination of the size of each groups territory and robust convergence. The usefulness of the proposed algorithm is verified by the application to a specially designed test function and a practical electromagnetic optimization problem.

Niching genetic algorithm adopting restricted competition selection combined with pattern search method

Among several niching methods, the restricted competition selection (RCS) method searches almost all the location of multiple solutions in the search space. This paper proposes a neo-niching method using the RCS method combined with pattern search method to reduce the calculation time. Through application to the numerical examples and optimal design of interior permanent magnet synchronous motor, it is verified that the proposed method searches the peaks fast and is appropriate for the design of electric machines.

Novel Memetic Algorithm implemented With GA (Genetic Algorithm) and MADS (Mesh Adaptive Direct Search) for Optimal Design of Electromagnetic System

This paper presents the novel implementation of the memetic algorithm with GA(Genetic Algorithm) and MADS(Mesh Adaptive Direct Search), which is applied for the optimal design methodology of the electric machine. This hybrid algorithm has been developed for obtaining the global optimum rapidly, which is effective for the optimal design of a electric machine with many local optima and much longer computation time. As a meta-heuristic search algorithm, MADS combined with a GA is validated with the Rastrigin function and the Shubert function with distinguished multimodal characteristics by investigating the evaluation number for optima convergence. In particular, the proposed algorithm has been forwarded to the optimal design of a direct-driven PM wind generator for maximizing the Annual Energy Production(AEP), of which design objective should be obtained by FEA(Finite Element Analysis). Finally, it is shown that MADS combined with GA has contributed to reducing the computation time effectively for the optimal design of a PM wind generator when compared with the purposely developed GA implemented with the parallel computing method.

Optimal design of permanent-magnet motor using autotuning niching genetic algorithm

A genetic algorithm using a niching method can be applied to a problem that requires multiple solutions in the search space. A novel niching genetic algorithm automatically tuning up the population size and niche radii is proposed. The proposed autotuning concept is embodied in a clearing method, elitism, and a deterministic method. Through application to the numerical examples and an optimal design of interior permanent-magnet synchronous motor, it is verified that the proposed method is appropriate for the design of electric machines.

Optimal Design of Direct-Driven PM Wind Generator for Maximum Annual Energy Production

Optimal design of the direct-driven PM wind generator, coupled with finite element analysis and genetic algorithm, has been performed to maximize the annual energy production over the whole wind speed characterized by the statistical model of wind speed distribution. Particularly, the parallel computing via internet web service has been applied to loose excessive computing times for optimization.

A Research on Iron Loss of IPMSM With a Fractional Number of Slot Per Pole

In this paper, we investigated rotor iron loss of interior permanent magnet synchronous machine (IPMSM), which has distributed armature windings. In order to study the iron loss with the effect of slot-pole combination, two machines for high-speed operation such as electric vehicle were designed. One is fractional slot winding and the other is conventional one. In the analysis, we developed a new iron loss model of electrical machines for high-speed operation. The calculated iron loss was compared with the experimental data. It was clarified that the proposed method can estimate iron loss effectively at high-speed operation. Based on this newly proposed method, we analyzed iron loss of the two machines according to their driving conditions. From the analysis results, it was shown that rotor iron loss of machine employing fractional slot-winding is definitely large at load condition. In addition, it is interesting that the ratio (rotor iron loss/stator iron loss) becomes larger as the speed of the machines increase if the number of slot per pole is fractional.

Harmonic Iron Loss Analysis of Electrical Machines for High-Speed Operation Considering Driving Condition

In this paper, we proposed new harmonic iron loss calculation methods for electrical machines. The new methods based on the concept of variable loss coefficients due to their flux density and frequency estimate harmonic iron loss in the time domain and the frequency domain. In order to verify the performance of the proposed method, we calculated the iron loss with conventional methods. Then, the estimated iron losses were compared with the experimental data. It was clarified that the proposed methods are more effective than the conventional methods at high-speed operation. Based on these newly proposed methods, we analyzed the iron loss of the machine for electric vehicle (EV) according to its driving condition. From the analysis results, it was shown that the harmonic iron losses of stator are larger than before at field-weakening region. In addition, it was revealed that rotor iron loss mainly induced by stator slot ripples does not show strong dependence on the current angle and only varied primarily according to the speed.