Showing papers by "Shun-Feng Su published in 2005"

Robust static output-feedback stabilization for nonlinear discrete-time systems with time delay via fuzzy control approach

Abstract: This paper addresses the problem of designing robust static output-feedback controllers for nonlinear discrete-time interval systems with time delays both in states and in control input. In the approach, we do not directly employ the Lyapunov approach, as do in most of traditional fuzzy control design approaches. Instead, sufficient conditions for guaranteeing the robust stability for the considered systems are derived in terms of the matrix spectral norm of the closed-loop fuzzy system. The stability conditions are further formulated into linear matrix inequalities so that the desired controller can be easily obtained by using the Matlab linear matrix inequality (LMI) toolbox. Finally, an example is provided to illustrate the effectiveness of the proposed approach.

Support vector regression performance analysis and systematic parameter selection

Abstract: Support vector regression (SVR) based on statistical learning is a useful tool for nonlinear regression problems. The SVR method deals with data in a high dimension space by using linear quadratic programming techniques. As a consequence, the regression result has optimal properties. However, if parameters were not properly selected, overfitting and/or underfilling phenomena might occur in SVR. Two parameters /spl sigma/, the width of Gaussian kernels and /spl epsi/, the tolerance zone in the cost function are considered in this research. We adopted the concept of the sampling theory into Gaussian filter to deal with parameter /spl sigma/. The idea is to analyze the frequency spectrum of training data and to select a cut-off frequency by including 90% of power in spectrum. The corresponding /spl sigma/ can then be obtained through the sampling theory. In our simulations, it can be found that good performances are observed when the selected frequency is near the cut-off frequency. For another parameter /spl epsi/, it is a tradeoff between the number of support vectors and the RMSE. By introducing the confidence interval concept, a suitable selection of /spl epsi/ can be obtained. The idea is to use the L/sub 1/-norm (i.e., when /spl epsi/ = 0 ) to estimate the noise distribution of training data. When /spl epsi/ is obtained by selecting the 90% confidence interval, simulations demonstrated superior performance in our illustrative example. By our systematical design, proper values of /spl sigma/ and /spl epsi/ can be obtained and the resultant system performances are nice in all aspects.

Dimension Reduction with Support Vector Regression for Ovarian Cancer Microarray Data

Abstract: In general, the support vector regression (SVR) is very suitable to approximate a high dimensionality space and ill-posed problem in modeling. That is, the SVR consists of a quadratic programming problem that can be solved efficiently and guaranteed to find a global extremism. Therefore, for the complex data, the SVR is easy to reconstruct an approximated model based on the linear programming technique. On the other hand, a typical microarray data consists of expression levels for a large number of genes on a relatively small number of samples. In order to avoid higher computational complexity and larger prediction errors on high-dimensional problem, we proposed the dimension reduction with SVR for the ovarian cancer microarray data. The SVR can reduce dimension on each sample from 9600 genes to about three hundreds genes. Besides, we can choose the epsiv value in the loss-function of SVR to obtain the variable number of gene and the proposed method can also overcome the block effect of microarray data. Finally, these results can provide for gene class discovery and gene class prediction

Ant colonies with cooperative process applied to the resource allocation problem

Abstract: In this paper, a novel algorithm describing ant colonies, with cooperation, is proposed to solve the resource allocation problem. The resource allocation problem is to allocate resources to activities, with the objectiveof optimizing the cost function. In our study, we viewed the search in ant colonies as a mechanism providing a main portion of diversity in search space. The cooperative process conducts fine-tuning for the solution provided by ant colonies, and it has the ability to escape from poor local optima. In this paper, several examples are tested to prove the superiority of our proposed algorithm. From simulation results, the proposed algorithm indeed has remarkable performance.

An Intelligent System for Multiple Sequences Alignment

Abstract: This paper provides an intelligent system, a fuzzy-genetic algorithm (FGA) with local search, for multiple sequences alignment. The general multiple sequence alignment, known as NP-hard problem, refers to search for maximal similarity in three or more sequences. The proposed algorithm is to enhance the performance of genetic algorithm by incorporating local search and fuzzy set theory for multiple sequence alignment. In the proposed algorithm, genetic algorithms perform a multiple directional search by maintaining a set of solutions. Local search operators are performed to explore the neighborhood in an attempt to enhance the fitness of the solution in a local manner. Moreover, fuzzy set theory is designed to dynamically adjust the probability of crossover, mutation and local search during evolutionary process. Results from our experiments indicate that our approach can obtain good performance in the majority of data sets with both low similarity and high diversity

A new convergence condition for discrete-time nonlinear system identification using a Hopfield neural network

Abstract: This paper presents a method of discrete time nonlinear system identification using a HopfieId neural network (HNN) as a coefficient learning mechanism to obtain optimized coefficients over a set of Gaussian basis functions. A linear combination of Gaussian basis functions is used to replace the nonlinear function of the equivalent discrete time nonlinear system. The outputs of the HNN, which are coefficients over a set of Gaussian basis functions, are discretized to be a discrete Hopfield learning model. Using the outputs of the HNN, one can obtain the optimized coefficients of the linear combination of Gaussian basis functions conditional on properly choosing an activation function scaling factor of the HNN. The main contributions of this paper is that the convergence of learning of the HNN can be guaranteed if the activation function scaling factor is properly chosen. Finally, to demonstrate the effectiveness of the proposed methods, simulation results are illustrated in this paper.