Osamu Takata

Fuzzy clustering of data with uncertainties using minimum and maximum distances based on L/sub 1/ metric

Fuzzy c-means is well-known among the various methods of fuzzy cluster analysis. L/sub 1/-based fuzzy c-means has also been studied in recent years. This paper discusses the L/sub 1/-based fuzzy c-means of data with fuzzy uncertainties. The data unit is supposed to be the Cartesian product of fuzzy numbers. The metric between a data unit with uncertainty and a cluster center is defined using minimum and maximum distances. The fuzzy c-means algorithm is an alternative procedure for the optimization of the cluster center and the fuzzy set membership, while the solution of the cluster center for uncertain data cannot be obtained directly. An algorithm for the solution of cluster centers based on the L/sub 1/ metric for uncertain data is developed in this paper. Using this algorithm, an exact alternate optimization procedure is obtained. Numerical examples show that the results for uncertain data are different from the results for data without uncertainties.

Knowledge-based system for die configuration design in cold forging

We propose a knowledge-based system for process planning in cold forging. In our study, we had developed FORTEK-L (FORging Technology of process planning using Expert Knowledge for Layout), which is widely applicable to various shapes and types of equipment. In the next step, we have developed a knowledge-based die configuration design system, FORTEK-D (Die configuration design), which can generate suitable die configurations from layouts obtained from FORTEK-L. In the development of FORTEK-D, we have analyzed and reproduced the expert engineer’s thinking process during die configuration design. Moreover, we have formulated the data representation, inference method, and knowledge-base for use with this system. In accordance with the formulation, we have developed a prototype system. We applied FORTEK-D to actual forged products and obtained appropriate results. The knowledge using the basic principle is widely applicable to various shapes and on a variety of equipment. Therefore, the abilities to maintain and operate the system have been improved.

Fuzzy c-Means and Mixture Distribution Model for Clustering Based on L1-Space

This paper aims at proposing and comparing two fuzzy models and a statistical model for clustering based on L1-space. Clustering methods in the fuzzy models are the standard fuzzy c-means and an entropy regularization method based on L1-space. Furthermore, we add new variables to them for improving the cluster division. In the statistical model, a mixture distribution model based on L1-space is proposed and the EM algorithm is applied.