• 한국지능시스템학회논문지
• /
• 제7권5호
• /
• pp.60-66
• /
• 1997

This paper presents a fuzzy inference-based reinforcement learning algorithm of dynamci recurrent neural networks, which is very similar to the psychological learning method of higher animals. By useing the fuzzy inference technique the linguistic and concetional expressions have an effect on the controller's action indirectly, which is shown in human's behavior. The intervlas of fuzzy membership functions are found optimally by genetic algorithms. And using recurrent neural networks composed of dynamic neurons as action-generation networks, past state as well as current state is considered to make an action in dynamical environment. We show the validity of the proposed learning algorithm by applying it to the inverted pendulum control problem.

• International Journal of Control, Automation, and Systems
• /
• 제1권2호
• /
• pp.194-202
• /
• 2003

In this paper, we introduce a category of Multi-FNN (Fuzzy-Neural Networks) models, analyze the underlying architectures and propose a comprehensive identification framework. The proposed Multi-FNNs dwell on a concept of fuzzy rule-based FNNs based on HCM clustering and evolutionary fuzzy granulation, and exploit linear inference being treated as a generic inference mechanism. By this nature, this FNN model is geared toward capturing relationships between information granules known as fuzzy sets. The form of the information granules themselves (in particular their distribution and a type of membership function) becomes an important design feature of the FNN model contributing to its structural as well as parametric optimization. The identification environment uses clustering techniques (Hard C - Means, HCM) and exploits genetic optimization as a vehicle of global optimization. The global optimization is augmented by more refined gradient-based learning mechanisms such as standard back-propagation. The HCM algorithm, whose role is to carry out preprocessing of the process data for system modeling, is utilized to determine the structure of Multi-FNNs. The detailed parameters of the Multi-FNN (such as apexes of membership functions, learning rates and momentum coefficients) are adjusted using genetic algorithms. An aggregate performance index with a weighting factor is proposed in order to achieve a sound balance between approximation and generalization (predictive) abilities of the model. To evaluate the performance of the proposed model, two numeric data sets are experimented with. One is the numerical data coming from a description of a certain nonlinear function and the other is NOx emission process data from a gas turbine power plant.

• 한국지능시스템학회:학술대회논문집
• /
• 한국퍼지및지능시스템학회 2005년도 추계학술대회 학술발표 논문집 제15권 제2호
• /
• pp.483-487
• /
• 2005

We introduce a new architecture of hetero-hybridized feed-forward neural networks composed of fuzzy set-based polynomial neural networks (FSPNN) and polynomial neural networks (PM) that are based on a genetically optimized multi-layer perceptron and develop their comprehensive design methodology involving mechanisms of genetic optimization and Information Granulation. The construction of Information Granulation based HFSPNN (IG-HFSPNN) exploits fundamental technologies of Computational Intelligence(Cl), namely fuzzy sets, neural networks, and genetic algorithms(GAs) and Information Granulation. The architecture of the resulting genetically optimized Information Granulation based HFSPNN (namely IG-gHFSPNN) results from a synergistic usage of the hybrid system generated by combining new fuzzy set based polynomial neurons (FPNs)-based Fuzzy Neural Networks(PM) with polynomial neurons (PNs)-based Polynomial Neural Networks(PM). The design of the conventional genetically optimized HFPNN exploits the extended Group Method of Data Handling(GMDH) with some essential parameters of the network being tuned by using Genetie Algorithms throughout the overall development process. However, the new proposed IG-HFSPNN adopts a new method called as Information Granulation to deal with Information Granules which are included in the real system, and a new type of fuzzy polynomial neuron called as fuzzy set based polynomial neuron. The performance of the IG-gHFPNN is quantified through experimentation.

• 한국지능시스템학회논문지
• /
• 제17권2호
• /
• pp.276-283
• /
• 2007

It has been shown that the structure identification of a neuro-fuzzy model improves their accuracy performances in a various modeling problems. In this paper, we claim that the structure identification of a neuro-fuzzy model can also reduce the degree of inconsistency of its fuzzy rulebase. Thus, the resulting neuro-fuzzy model serves as more like a structured knowledge representation scheme. For this, we briefly review a structure identification method of a neuro-fuzzy model and propose a systematic method to measure inconsistency of a fuzzy rulebase. The proposed method is applied to problems or fuzzy system reproduction and nonlinear system modeling in order to validate our claim.

• Journal of applied mathematics & informatics
• /
• 제15권1_2호
• /
• pp.265-275
• /
• 2004

In this paper, multi-objective models for designing 3D trajectory of horizontal wells are developed in a fuzzy environment. Here, the objectives of minimizing the length of the trajectory and the error of entry target point are fuzzy in nature. Some parameters, such as initial value, end value, lower bound and upper bound of the curvature radius, tool-face angle and the arc length of each curve section, are also assumed to be vague and imprecise. The impreciseness in the above objectives have been expressed by fuzzy linear membership functions and that in the above parameters by triangular fuzzy numbers. Models have been solved by the fuzzy non-linear programming method based on Zimmermann [1] and Lee and Li [2]. Models are applied to practical design of the horizontal wells. Numerical results illustrate the accuracy and efficiency of the fuzzy models.

• 한국지능시스템학회논문지
• /
• 제14권3호
• /
• pp.324-329
• /
• 2004

에이전트에서 수행할 수 있는 작업들에 대한 처리시간이 실제 작업 전에는 퍼지값으로만 주어지고, 실제 작업이 수행될 때야 작업 시간이 결정되는 다중 에이전트 시스템에 대해서 작업을 에이전트들에 할당하고 스케쥴링하는 작업조정 방법을 제안한다. 제안한 방법은 두 단계의 유전자 알고리즘으로 구성되는데, 상위 단계의 유전자 알고리즘에서는 작업들을 적합한 에이전트에 할당하는 역할을 하고, 하위 단계의 유전자 알고리즘은 첫 번째 유전자 알고리즘의 제시하는 작업 할당 방법에 가장 적합한 작업 스케쥴을 탐색하는 역할을 한다. 이 논문에서는 제안한 유전자 알고리즘 기반 작업 조정 방법을 소개한 다음, 에이전트가 고장 등으로 동작할 수 있는 장애가 발생할 때 처리하는 기법을 소개하고, 제안한 방법을 구현하여 실험한 결과를 보인다.

• 대한전기학회:학술대회논문집
• /
• 대한전기학회 2001년도 하계학술대회 논문집 D
• /
• pp.1958-1960
• /
• 2001

In this paper, we propose the design methodology of target tracking system using the identification of TS fuzzy model based on genetic algorithm(GA) and RLS algorithm. In general, the objective of target tracking is to estimate the future trajectory of the target based on the past position of the target obtained from the sensor. In the conventional and mathematical nonlinear filtering method such as extended Kalman filter(EKF), the performance of the system may be deteriorated in highly nonlinear situation. In this paper, to resolve these problems of nonlinear filtering technique, the error of EKF by nonlinearity is compensated by identifying TS fuzzy model. In the proposed method, after composing training datum from the parameters of EKF, by identifying the premise and consequent parameters and the rule numbers of TS fuzzy model using GA, and by tuning finely the consequent parameters of TS fuzzy model using recursive least square(RLS) algorithm, the error of EKF is compensated. Finally, the proposed method is applied to three dimensional tracking problem, and the simulation results shows that the tracking performance is improved by the proposed method.

• 한국경영과학회:학술대회논문집
• /
• 대한산업공학회/한국경영과학회 2002년도 춘계공동학술대회
• /
• pp.877-884
• /
• 2002

This paper deals with a multi-stage TSK fuzzy modeling method by using Genetic Programming (GP). Based on the time sequence of sampling data the best structural change points of complex systems are detemined by using GP, and also the moving window is simultaneously introduced to overcome the excessive amount of calculation during the generating procedure of GP tree. Therefore, a multi-stage TSK fuzzy model that attempts to represent a complex problem by decomposing it into multi-stage sub-problems is addressed and its learning algorithm is proposed based on the Radial Basis Function (RBF) network. This approach allows us to determine the model structure and parameters by stages so that the problems ran be simplified.

• 한국지능시스템학회논문지
• /
• 제10권6호
• /
• pp.575-583
• /
• 2000

본 논문에서 임의의 데이터가 입력되면 기준 영상 중에서 가장 유사도가 큰 영상을 찾아 국부 승리자로 선택하고, 그 국부 승리자 중에서 전체 승리자를 선택하여 최종 출력값을 얻는 계층적 FGNN(Fuzzy Genetic Neural Network)을 제안하고, 이에 하이브리드 퍼지 소속함수와 유전자 알고리즘을 적용하였다. 하이브리드 퍼지 소속함수는 입력 값을 0~1 사이의 값으로 함으로써 시스템의 속도를 빠르게 하고 유전자 알고리즘을 입력값을 일정한 오차 이내로 하여 최적의 영상을 얻도록 하였다. 위의 계층적 FGNN 알고리즘을 회로 설계 및 검증하였다. 또한 제안한 FGNN을 이용하여 영상에 포함된 잡음을 제거하고, 이와 유사한 구조를 가진 FDNN(Fuzzy Decision Neural Network) 성능보다 FGNN의 성능이 우수함을 여러 가지 영상을 통하여 확인하였다. 또한 모의 실험 결과 영상에 대한 평균자승오차(MSE : Mean Square Error)를 비교하였으며, 그 결과 하이브리드 퍼지 함수와 유전자 알고리즘을 적용한 FGNN이 메디안 필터, OC, CO, FDNN 등에 비해 우수함을 확인하였다. FGNN 알고리즘을 Top-Down 방식으로 VHDL(VHSIC Hardware description Language)을 이용하여 코딩(Coding)하고, Synopsys 툴을 이용하여 하드웨어를 설계하였다. 이 알고리즘의 하드웨어는 총 5개의 블록으로 가지고 있고 각각의 블록은 파이프라인 형태로 구성하고, 이는 Synopsys 툴을 이용하여 동작 및 성능을 검증하였다.

• 대한전기학회:학술대회논문집
• /
• 대한전기학회 1999년도 하계학술대회 논문집 B
• /
• pp.501-503
• /
• 1999

In this paper, the design method of a hybrid fuzzy controller with an optimal auto-tuning method is proposed. The conventional PID controller becomes so sensitive to the control environments and the change of parameters that the efficiency of its utility for the complex and nonlinear plant has been questioned in transient state. In this paper, first, a hybrid fuzzy logic controller(HFLC) is proposed. The control input of the system in the HFLC is a convex combination by a fuzzy variable of the FLC's output in transient state and the PID's output in steady state. Second, a powerful auto-tuning algorithm is presented to automatically improve the Performance of controller, utilizing the improved complex method and the genetic algorithm. The algorithm estimates automatically the optimal values of scaling factors and PID coefficients. Controllers are applied to the plants with time-delay and the DC servo motor Computer simulations are conducted at the step input and the system performances are evaluated in the ITAE.