Showing papers in "International Journal of Applied Mathematics and Computer Science in 2000"

Fusion Technology of Neural Networks and Fuzzy Systems: A Chronicled Progression from the Laboratory to Our Daily Lives

Abstract: We chronicle the research on the fusion technology of neural networks and fuzzy systems (NN+FS), the models that have been proposed from this research, and the commercial products and industrial systems that have adopted these models. First, we review the NN+FS research activity during the early stages in Japan, the US, and Europe. Next, following the classification of NN+FS models, we show the ease of fusing these technologies based on the similarities of the data flow network structures and the non-linearity realization strategies of NNs and FSs. Then, we describe several models and applications of NN+FS. Finally, we introduce some important and recently developed NN+FS patents. 1I ntroduction During the late 1980s, the number of researchers and engineers interested in neural networks (NNs) and fuzzy logic (FL) increased, dramatically introducing the NN and FL technologies into several application fields. Both technologies are widely used and are considered fundamental engineering technologies. As the two technologies developed at same time, researchers and engineers studied the technologies’ similarities and complementarities and began developing a fusion model. Although the first paper that used both NNs and fuzzy sets as keywords can be found in 1974, research on fusing NNs and fuzzy systems (FSs) essentially began in 1988 and has dramatically increased since then. Within several years, NN+FS fusing technology was already being used in commercial products and industrial systems. The practicality of technology, introduced at the beginning of this paper, is supported by the number of real-word applications based on this technology. Since its introduction, the fusion technology has widely expanded into application fields requiring such tasks as control, operations research, retrieval, clustering, speech recognition, and others [15]. We show the spread of this trend from Japan to the US and Europe in next section. Figure 1 shows the increasing number of papers and conferences that using both the keywords NN and FS during the early 1990s. Furthermore, genetic algorithms (GA) were introduced to the fusing technologies and several combinations of GA and FS and of GA and NN have been proposed since 1989. Consumer products that use these GA+NN and GA+FS cooperative models have also been put on the market. In this paper, we describe the fusing technologies of Soft Computing from a historical and RD we especially focus on NN+FS research models and application tasks in Japan where they were primarily researched and developed. For the past 10 years, fundamental patents on NN+FS have been established. We introduce some of these patents in section 5.

Variable- and Fixed-Structure Augmented Interacting Multiple Model Algorithms for Manoeuvring Ship Tracking Based on New Ship Models

Abstract: The real-world tracking applications meet a number of difficulties caused by the presence of different kinds of uncertainty - unknown or not precisely known system model and random processes’ statistics or due to abrupt changes in the system modes of functioning. These problems are especially complicated in the marine navigation practice, where the commonly used simple models of rectilinear or curvilinear target motions do not match to the highly non-linear dynamics of the manoeuvring ship motion. A solution of these problems is to derive more adequate descriptions of the real ship dynamics and to design adaptive estimation algorithms. After analysis of basic hydrodynamic models, new ship models are derived in the paper. They are implemented in two versions of the recently very popular Interacting Multiple Model (IMM) algorithm. The first one is a standard IMM version using preliminary defined fixed structure (FS) of models. They represent various modes of ship motion, distinguished by their rate of turns. The same rate of turn is additionally adjusted in the proposed new augmented versions of the IMM (AIMM) algorithm by using FS and variable structure (VS) of adaptive models estimating the current change of the system control parameters. The obtained Monte Carlo simulation results show that the VS AIMM algorithm outperforms the FS AIMM and FS IMM algorithms with respect to accuracy and adaptability.

Classification, Association and Pattern Completion Using Neural Similarity Based Methods

Abstract: A framework for Similarity-Based Methods (SBMs) includes many classification models as special cases: neural network of the Radial Basis Function Networks type, Feature Space Mapping neurofuzzy networks based on separable transfer functions, Learning Vector Quantization, variants of the k nearest neighbor methods and several new models that may be presented in a network form Multilayer Perceptrons (MLPs) use scalar products to compute weighted activation of neurons, combining soft hyperplanes to provide decision borders Distance-based multilayer perceptrons (D-MLPs) evaluate similarity of inputs to weights offering a natural generalization of standard MLPs Cluster-based initialization procedure determining architecture and values of all adaptive parameters is described Networks implementing SBM methods are useful not only for classification and approximation, but also as associative memories, in problems requiring pattern completion, offering an efficient way to deal with missing values NonEuclidean distance functions may also be introduced by normalization of the input vectors in an extended feature space Both approaches influence shapes of decision borders dramatically An illustrative example showing these changes is provided

Existence and Computation of the Set of Positive Solutions to Polynomial Matrix Equations

Abstract: Abstract Necessary and sufficient conditions are established for the existence of positive solutions to polynomial diophatine equations. A method for computation of the set of positive solutions to polynomial diophatine equation based on extreme points and extreme directions is proposed. The effectiveness of the method is demonstrated on a numerical example.

Sustainable exploitation of biological populations

Abstract: For a system of nonlinear difference equations that models the dynamics of exploited biological populations, a locally optimal periodic solutionis constructed. If this solution is unstable, a stabilizing feedback in the harvesting is introduced. The method is applied to an age-structured population model in fishery as well as to a host-parasitoid system for which the number of hosts and the number of introduced parasitoids shouldbe minimized.