• 제어로봇시스템학회:학술대회논문집
• /
• 2005.06a
• /
• pp.1080-1085
• /
• 2005

A hybrid control scheme is proposed for the stabilization of backward movement along simple paths for a vehicle composed of a truck and six trailers. The hybrid comprises the combination of a linear quadratic regulator (LQR) and a neurocontroller (NC) that is trained by a genetic algorithm (GA). Acting singly, either the NC or the LQR are unable to perform satisfactorily over the entire range of the operation required, but the proposed hybrid is shown to be capable of providing good overall system performance. The evaluation function of the NC in the hybrid design has been modified from the conventional type to incorporate both the squared errors and the running steps errors. The reverse movement of the trailer-truck system can be modeled as an unstable nonlinear system, with the control problem focusing on the steering angle. Achieving good backward movement is difficult because of the restraints of physical angular limitations. Due to these constraints the system is impossible to globally stabilize with standard smooth control techniques, since some initial states necessarily lead to jack-knife locks. This paper demonstrates that a hybrid of neural networks and LQR can be used effectively for the control of nonlinear dynamical systems. Results from simulated trials are reported.

• The Transactions of the Korea Information Processing Society
• /
• v.6 no.2
• /
• pp.439-448
• /
• 1999

In this paper, two optimization algorithms based on artificial neural networks and genetic algorithms are proposed for cellular radio channel assignment problems. The channel assignment process is characterized as minimization of the energy function which represents constraints of the channel assignment problems. All three constraints such as the co-channel constraint, the adjacent channel constraint and the co-site channel constraint are considered. In the neural networks approach, certain techniques such as the forced assignment and the changing cell order are developed, and in the genetic algorithms approach, data structure and proper genetic operators are developed to find optimal solutions, As simulation results, the convergence rates of the two approaches are presented and compared.

• Journal of the Korean Operations Research and Management Science Society
• /
• v.25 no.4
• /
• pp.97-109
• /
• 2000

This work proposes a neural network approach to solve vehicle routing problems which have diverse application areas such as vehicle routing and robot programming. In solving these problems, classical mathematical approaches have many difficulties. In particular, it is almost impossible to implement a real-time vehicle routing with multiple vehicles. Recently, many researchers proposed methods to overcome the limitation by adopting heuristic algorithms, genetic algorithms, neural network techniques and others. The most basic model for path planning is the Travelling Salesman Problem(TSP) for a minimum distance path. We extend this for a problem with dynamic upcoming of new positions with multiple vehicles. In this paper, we propose an algorithm based on SOM(Self-Organization Map) to obtain a sub-optimal solution for a real-time vehicle routing problem. We develope a model of a generalized multiple TSP and suggest and efficient solving procedure.

• Proceedings of the KIEE Conference
• /
• 1999.07c
• /
• pp.1177-1179
• /
• 1999

This paper presents a design of artificial neural network power system stabilizer(ANNPSS) using adaptive evolutionary algorithm(AEA). We have proposed an adaptive evolutionary algorithm which uses both a genetic algorithm(GA) and an evolution strategy(ES), useing the merits of two different evolutionary computations. ANNPSS shows better control performances than conventional power system stabilizer(CPSS) in three-phase fault with heavy load which is used when tuning ANNPSS. To show the robustness of the proposed ANNPSS, it is applied to damp the low frequency oscillation caused by disturbances such as three-phase fault with normal and light load. the proposed ANNPSS shows better robustness than CPSS.

• Proceedings of the Korean Institute of Electrical and Electronic Material Engineers Conference
• /
• 2010.06a
• /
• pp.206-206
• /
• 2010

본 연구에서는 신경망을 이용하여 SiN 박막의 특성을 예측하는 모델을 개발하였다. 신경망으로는 일반화된 회귀 신경망 (generalized regression neural network-GRNN)을 이용하였고, GRNN 모델의 예측수행은 유전자 알고리즘 (genetic algorithm-GA)을 이용하여 최적화 하였다. 개발된 모델을 이용하여 증착률과 굴절률 및 균일도를 공정변수의 함수로 예측하였다.

• International Journal of Fuzzy Logic and Intelligent Systems
• /
• v.3 no.1
• /
• pp.78-86
• /
• 2003

Multivariable control system exist widely in many types of systems such as chemical processes, biomedical processes, and the main steam temperature control system of the thermal power plant. Up to the present time, Pill Controllers have been used to operate these systems. However, it is very difficult to achieve an optimal PID gain with no experience, because of the interaction between loops and gain of the Pill controller has to be manually tuned by trial and error. This paper suggests a tuning method of the Pill Controller for the multivariable power plant using an immune algorithm, through computer simulation. Tuning results by immune algorithms based neural network are compared with the results of genetic algorithm.

• Proceedings of the Korean Institute of Intelligent Systems Conference
• /
• 2005.04a
• /
• pp.297-300
• /
• 2005

In this study, we proposed genetically optimized self-organizing fuzzy polynomial neural network based on information granulation and evolutionary algorithm (gdSOFPNN), develop a comprehensive design methodology involving mechanisms of genetic optimization. The proposed gdSOFPNN gives rise to a structural Iy and parametrically optimized network through an optimal parameters design available within FPN (viz. the number of input variables, the order of the polynomial, input variables, the number of membership functions, and the apexes of membership function). Here, with the aid of the information granulation, we determine the initial location (apexes) of membership functions and initial values of polynomial function being used in the premised and consequence part of the fuzzy rules respectively. The performance of the proposed gdSOFPNN is quantified through experimentation that exploits standard data already used in fuzzy modeling.

• The Transactions of the Korean Institute of Electrical Engineers D
• /
• v.53 no.6
• /
• pp.433-439
• /
• 2004

The Helicopter Simulator System is non-linear and complex. Futhermore, because of absence of its accurate mathematical model, it is difficult to control accurately its attitudes such as elevation angle and azimuth one. Therefore, we proposed a Hybrid GA-PID WAVENET(Genetic Algorithm Proportional Integral Derivative Wavelet Neural Network)control technique to control efficiently these angles. The proposed Hybrid GA-PID WAVENET is made through the following process. First, the WAVENET fundamental functions are defined. And their dilation and translation values are adjusted by GA to construct the optimal WAVENET controller. Secondly, the proportional, integral, and derivative gain coefficients of PR controller are tuned optimally. Finally, WAVENET controller which has a good transient characteristic and GA-PE controller which has a good steady state characteristic is adequately combined in hybrid type. Through the computer simulations, it is proved that the Hybrid GA-PE WAVENET control technique has a more excellent dynamic response than PID control technique and GA-PID one.

• Proceedings of the Korean Institute of Electrical and Electronic Material Engineers Conference
• /
• 2007.11a
• /
• pp.111-112
• /
• 2007

Silicon nitride (SiN) 박막을 플라즈마 응용화학기상법을 이용하여 증착하였다. SiN박막의 전하밀도는 일반화된 회귀 신경망과 유전자 알고리즘을 이용하여 모델링하였다. PECVD 공정은 Box Wilson 실험계획표를 이용하여 수행하였다. $SiH_4$ 유량변화에 따른 온도의 영향은 미미하였다. 그러나, 저 전력에서의 온도증가 (또는 저온에서의 전력의 증가)에 따라 전하밀도는 급격히 상승하였으며, 이는 [N-H]의 증가에 기인하는 것으로 해석되었다. $SiH_4$ 유량의 증가 (또는 고온에서의 전력의 증가)에 따라 전하밀도는 감소하고 있으며, 이는 [Si-H]의 증가에 기인하는 것으로 이해된다.

• The Transactions of The Korean Institute of Electrical Engineers
• /
• v.67 no.12
• /
• pp.1665-1670
• /
• 2018

A convolutional neural network (CNN), which is one of the deep learning models, has been very successful in a variety of computer vision tasks. Filters of a CNN are automatically generated, however, they can be further optimized since there exist the possibility of existing redundant and less important features. Therefore, the aim of this paper is a filter reduction to accelerate and compress CNN models. Evolutionary algorithms is adopted to remove the unnecessary filters in order to minimize the parameters of CNN networks while maintaining a good performance of classification. We demonstrate the proposed filter reduction methods performing experiments on CIFAR10 data based on the classification performance. The comparison for three approaches is analysed and the outlook for the potential next steps is suggested.