Browsing by Author "Cheng, Yi-Chang"

Now showing 1 - 4 of 4

A Novel Self-Constructing Evolution Algorithm for TSK-type Fuzzy Model Design
(2010)
Lin, Sheng-Fuu
;
Chang, Jyun-Wei
;
Cheng, Yi-Chang
;
Hsu, Yung-Chi
;
電機工程學系
In this paper, a novel self-constructing evolution algorithm (SCEA) for TSK-type fuzzy model (TFM) design is proposed. The proposed SCEA method is different from the traditional genetic algorithms (GA). A chromosome of the population in GA represents a full solution and only one population presents all solutions. Our method applies a population to evaluate a partial solution locally, and several populations to construct the full solution. Thus, a chromosome represents only partial solution. The proposed SCEA uses the self-constructing learning algorithm to construct the TFM automatically that is based on the input data to decide the input partition. And we also adopted the sequence search-based dynamic evolution (SSDE) method to perform parameter learning. Simulation results have shown that the proposed SCEA method obtains better performance than some existing models.
Multi groups cooperation based symbiotic evolution for TSK-type neuro-fuzzy systems design
(2010-07-01)
Hsu, Yung-Chi
;
Lin, Sheng-Fuu
;
Cheng, Yi-Chang
;
電控工程研究所
;
Institute of Electrical and Control Engineering
In this paper, a TSK-type neuro-fuzzy system with multi groups cooperation based symbiotic evolution method (TNFS-MGCSE) is proposed. The TNFS-MGCSE is developed from symbiotic evolution. The symbiotic evolution is different from traditional GAs (genetic algorithms) that each chromosome in symbiotic evolution represents a rule of fuzzy model. The MGCSE is different from the traditional symbiotic evolution; with a population in MGCSE is divided into several groups. Each group formed by a set of chromosomes represents a fuzzy rule and cooperate with other groups to generate the better chromosomes by using the proposed cooperation based crossover strategy (CCS). In this paper, the proposed TNFS-MGCSE is used to evaluate by numerical examples (Mackey-Glass chaotic time series and sunspot number forecasting). The performance of the TNFS-MGCSE achieves excellently with other existing models in the simulations. (C) 2010 Elsevier Ltd. All rights reserved.
Precise image alignment using cooperative neural-fuzzy networks with association rule mining-based evolutionary learning algorithm
(2012-02-01)
Hsu, Chi-Yao
;
Cheng, Yi-Chang
;
Lin, Sheng-Fuu
;
電機工程學系
;
Department of Electrical and Computer Engineering
Precise image alignment is considered a critical issue in industrial visual inspection, since it performs an accurate pose to the object in inspected images. Recently, image alignment based on neural networks has become very popular due to its performance at speed. However, such a method has difficulty when applied to the alignment of images on a large range of affine transformation. To address this, a cooperative neural-fuzzy network (CNFN) with association rule mining-based evolutionary learning algorithm (ARMELA) is proposed. Unlike traditional neural network-based approaches, the proposed CNFN utilizes a coarse-to-fine alignment procedure to adapt image alignment to a larger range of affine transformation. The proposed ARMELA combines the self-adaptive method and association rules selection method to self-adjust the structure and parameters of the neural-fuzzy network. Furthermore, L2 regularization is adopted to control ARMELA such that the convergence speed increases. Experimental results show that the performance of the proposed scheme is superior to the traditional neural network methods in terms of accuracy and robustness. (c) 2012 Society of Photo-Optical Instrumentation Engineers (SPIE). [DOI: 10.1117/1.OE.51.2.027006]
TWO-STRATEGY REINFORCEMENT EVOLUTIONARY ALGORITHM USING DATA-MINING BASED CROSSOVER STRATEGY WITH TSK-TYPE FUZZY CONTROLLERS
(2010-09-01)
Lin, Sheng-Fuu
;
Cheng, Yi-Chang
;
電機工程學系
;
Department of Electrical and Computer Engineering
This paper proposes a two-strategy reinforcement evolutionary algorithm using data-mining crossover strategy (TSR-EADCS) with a TSK-type fuzzy controller (TFC) for solving various control problems. The purpose of the R-EA DCS is not only to improve the design of traditional reinforcement signal but also to determine the suitable rules in a TEC and the suitable groups that are selected to perform crossover operation. Therefore, this paper proposes a two-strategy reinforcement signal to improve the performance of the traditional reinforcement signal design and uses the data mining technique to find suitable fuzzy rules and groups for evolution. The proposed TSR-EADCS consists of both structure and parameter learning. Ins fracture learning, the TSR-EADCS uses the self adaptive method to determine the suitability of TEC models between different numbers of fuzzy rules. In parameter learning, the TSR-EADCS uses the data-mining crossover strategy which is based on frequent pattern growth to select the suitable groups that are used to perform crossover operation. Illustrative examples are conducted to show the performance and applicability of the TSR-EADCS.