-
[in Japanese]
Article type: Article
1999 Volume 11 Issue 3 Pages
359-
Published: June 15, 1999
Released on J-STAGE: January 08, 2018
JOURNAL
FREE ACCESS
-
Yasuhito MUKOUCHI
Article type: Article
1999 Volume 11 Issue 3 Pages
360-368
Published: June 15, 1999
Released on J-STAGE: January 08, 2018
JOURNAL
FREE ACCESS
-
Kyoichi KIJIMA
Article type: Article
1999 Volume 11 Issue 3 Pages
369-381
Published: June 15, 1999
Released on J-STAGE: January 08, 2018
JOURNAL
FREE ACCESS
-
Kazumasa SHINJO
Article type: Article
1999 Volume 11 Issue 3 Pages
382-395
Published: June 15, 1999
Released on J-STAGE: January 08, 2018
JOURNAL
FREE ACCESS
-
Junji NISHINO
Article type: Article
1999 Volume 11 Issue 3 Pages
396-402
Published: June 15, 1999
Released on J-STAGE: January 08, 2018
JOURNAL
FREE ACCESS
-
Kenji AOYAMA
Article type: Article
1999 Volume 11 Issue 3 Pages
403-407
Published: June 15, 1999
Released on J-STAGE: January 08, 2018
JOURNAL
FREE ACCESS
-
Article type: Bibliography
1999 Volume 11 Issue 3 Pages
408-413
Published: June 15, 1999
Released on J-STAGE: January 08, 2018
JOURNAL
FREE ACCESS
-
Kazuo NAKAMURA
Article type: Article
1999 Volume 11 Issue 3 Pages
414-415
Published: June 15, 1999
Released on J-STAGE: January 08, 2018
JOURNAL
FREE ACCESS
-
[in Japanese]
Article type: Article
1999 Volume 11 Issue 3 Pages
416-
Published: June 15, 1999
Released on J-STAGE: January 08, 2018
JOURNAL
FREE ACCESS
-
[in Japanese]
Article type: Article
1999 Volume 11 Issue 3 Pages
416-
Published: June 15, 1999
Released on J-STAGE: January 08, 2018
JOURNAL
FREE ACCESS
-
[in Japanese]
Article type: Article
1999 Volume 11 Issue 3 Pages
417-
Published: June 15, 1999
Released on J-STAGE: January 08, 2018
JOURNAL
FREE ACCESS
-
Gengui Zhou
Article type: Article
1999 Volume 11 Issue 3 Pages
418-
Published: June 15, 1999
Released on J-STAGE: January 08, 2018
JOURNAL
FREE ACCESS
-
Yinzhen Li
Article type: Article
1999 Volume 11 Issue 3 Pages
418-
Published: June 15, 1999
Released on J-STAGE: January 08, 2018
JOURNAL
FREE ACCESS
-
Makoto NISHIDA, Takako SATO, Chikako ISHIZAWA
Article type: Article
1999 Volume 11 Issue 3 Pages
419-428
Published: June 15, 1999
Released on J-STAGE: January 08, 2018
JOURNAL
FREE ACCESS
Many pieces of tradition on weather predictions are seen all over Japan and they have customarily been utilized based on people's experience. The verbally expressed predictions based on the tradition are usually vague, and scientific analysis have not been attempted so far to the prediction accuracy of them. Results concerning the test of the temperature prediction accuracy are presented in this paper. The temperature prediction in a district is one of the constituents of long-term weather prediction based on an empirical "kan-dameshi" or "a prediction derived from cold season temperature variations". In actual procedure, numeric data which are the translations of the verbally expressed temperature predictions are used for the test by employing a coincidence rate and a score method. Another test concerning the temperature prediction is practiced as well with the verbally expressed every-three-months forecast announced by the Meteorological Agency. Our results over one year by using five-temperature-classes score method about "kan-dameshi" represented, on average, nearly 53% prediction accuracy, which is as same as the accuracy obtained from the Meteorological Agency data.
View full abstract
-
Masatoshi SAKAWA, Toshihiro SHIBANO, Kosuke KATO, Shinji MIYAHARA
Article type: Article
1999 Volume 11 Issue 3 Pages
429-437
Published: June 15, 1999
Released on J-STAGE: January 08, 2018
JOURNAL
FREE ACCESS
In this paper, we focus on multiobjective intger programming problems involving fuzzy numbers to reflect the experts' ambiguous understanding of the nature of the parameters in the problem-formulation process. For such multiobjective integer programming problems with fuzzy parameters, considering the imprecise nature of the decision maker's judgements, we propose an interactive fuzzy satisficing method through genetic algorithms for deriving a satisficing solution of the decision maker. In the genetic algorithms, triple string representation and random arithmetic crossover are proposed for the purpose of the efficient search. A lot of illustrutive numerical examples demonstrate the effectiveness of the proposed method.
View full abstract
-
Manabu NII, Kimiko TANAKA, Hisao ISHIBUCHI
Article type: Article
1999 Volume 11 Issue 3 Pages
438-452
Published: June 15, 1999
Released on J-STAGE: January 08, 2018
JOURNAL
FREE ACCESS
When a fuzzy vector is presented to a multi-layer feedforward neural network, the calculation of the corresponding fuzzy output vector is performed by fuzzy arithmetic. Since fuzzy arithmetic is independently executed at each unit of the neural network, the fuzziness of the fuzzy output from each unit is increased by the feedforward calculation in the neural network. In this paper, we first discuss such increase of the fuzziness from a viewpoint of the local application of the extension principle. In the local application, the extension principle is locally applied to the fuzzy input-output relation of each unit. On the other hand, theoretically the extension principle can be globally applied to the fuzzy input-output relation of the whole neural network. The global application is, however, very difficult in practice. Next we illustrate the increase of the fuzziness in the local application of the extension principle by computer simulations on numerical examples. Then we describe subdivision methods applied to the level sets of fuzzy input vectors for avoiding the increase of the fuzziness in fuzzy arithmetic. Finally we examine the effectiveness of the subdivision methods by applying them to rule extraction problems.
View full abstract
-
Kazuya KISHIDA, Shinya FUKUMOTO, Hiromi MIYAJIMA
Article type: Article
1999 Volume 11 Issue 3 Pages
453-461
Published: June 15, 1999
Released on J-STAGE: January 08, 2018
JOURNAL
FREE ACCESS
There are several fuzzy models using self-organization and vector quantization. It is well known that these models effectively construct fuzzy rules representing the distribution of input data, and are not affected even when the number of input dimensions increases. However most of these models are given the number of fuzzy rules in advance. In this paper, fuzzy rules are created sequentially so as to satisfy an objective value, and the proper number of them is determined finally. That is, the number of neurons which are reference vectors, is determined by using a self-creating neural network first. From the result, fuzzy rules are determined by using the descent method. Then if input-output data are approximated so as to satisfy the objective value, an algorithm terminates. Otherwise an algorithm for increasing the number of neurons is repeated. In order to show the validity of the proposed method, we performed some numerical examples.
View full abstract
-
Taro FUJIMOTO, Michio SUGENO
Article type: Article
1999 Volume 11 Issue 3 Pages
462-471
Published: June 15, 1999
Released on J-STAGE: January 08, 2018
JOURNAL
FREE ACCESS
Thesaurus is necessary for natural language processing, however, the construction of the thesaurus requires extreme costs. Hence, several automatic thesaurus construction methods are proposed to reduce these costs. One typical method uses mutual information and this method can construct admissible capability thesaurus, however, it is so complex that it needs large calculation cost. In this paper, we propose low cost and accurate method using fuzzy proximity relation and fuzzy filters. We use the fuzzy proximity relation to define proximity between verbs, two types of fuzzy filters to reflect appearance information. One fuzzy filter reduces the effect of low frequency verbs and the other reduces the effect of high frequency nouns. We combine two types of filters and fuzzy proximity relations and define four types of fuzzy proximity. To evaluate these proposal proximity, we use mutual information to define proximity between verbs, and combine two types of fuzzy filters with this proximity. Five types of frequency data characterized by post-position WO, NI, HA, GA, DE are extracted form EDR co-occurrence dictionary as a trainig data. First, we re-classify the verbs that are entries of BUNRUI GOI HYOU(a Japanese thesaurus)to evaluate the eight proximity. Second, we classify verbs which are not entries of BUNRUI GOI HYOU using these eight proximity and evaluate by person. The result of these two evaluation is that the proximity using fuzzy proximity relation with fuzzy filter which reduces the effect of low frequency verbs using the data including the post-position DE gets the highest evaluation. And this fuzzy proximity reduces 7% of calculation time than the proximity using mutual information method.
View full abstract
-
Masahiro INUIGUCHI, Tetsuzo TANINO
Article type: Article
1999 Volume 11 Issue 3 Pages
472-481
Published: June 15, 1999
Released on J-STAGE: January 08, 2018
JOURNAL
FREE ACCESS
In this paper, we develop a DEA (Data Envelopment Analysis)with fuzzy input-output data. There are several approaches to extend the DEA into the case of fuzzy input-output data. We made two natural extension approaches. In the first approach, a linear programming problem solved in the conventional DEA is regarded as a mapping from an input-output data set to a D-efficiency value set. Applying the extension principle to the mapping, we obtain a fuzzy set of D-efficiency values from given fuzzy input-output data. In the second approach, we propose an efficiency analysis based on possibility theory. By this approach, possible and necessary efficiencies are defined. We discuss improvement guides for a DMU (Decision Making Unit) to be possibly and necessarily efficient to a given degree. The relations between the fuzzy set of D-efficiency values and the possibilistic efficiency analysis are investigated. A numerical example is given to illustrate the proposed analyses.
View full abstract
-
Masatoshi SAKAWA, Ichiro Nishizaki, Yoshio UEMURA, Keiichi KUBOTA
Article type: Article
1999 Volume 11 Issue 3 Pages
482-491
Published: June 15, 1999
Released on J-STAGE: January 08, 2018
JOURNAL
FREE ACCESS
This paper persents interactive fuzzy programming for a decentralized two-level linear programming problem with one decision maker (DM) at the upper level and multiple DMs at the lower level. The interactive method consists of two phase. In the first phase, after determining the fuzzy goals of the DM at both levels, the DM at the upper level subjectively specifies a minimal satisfactory level. Taking into consideration overall satisfactory balance between the two levels, the DM at the upper level updates the minimal satisfactory level if necessary. In the second phase, consulting the ratios of satisfaction between the DM at the upper level and each of the DMs at the lower level, the DM at the upper level specifies maximal permissible satisfactory levels to the DMs at the lower level and updates them if necessary. A satisfactory solution can be derived efficiently through the interactive procedure of the two phases. Moreover, the derived satisfactory solution is an M-Pareto optimal solution in the satisfactory space.
View full abstract
-
Jianqiang YI, Naoyoshi YUBAZAKI, Kaoru HIROTA
Article type: Article
1999 Volume 11 Issue 3 Pages
492-502
Published: June 15, 1999
Released on J-STAGE: January 08, 2018
JOURNAL
FREE ACCESS
SIRMs(Single Input Rule Modules)dynamically connected fuzzy inference model is applied to inverted pendulum and cart systems, and is shown to be effective even though the control objects have 4 input items. A fuzzy controller for stabilizing inverted pendulum and cart systems is constructed by taking the control features of the objects into consideration in the design of the SIRMs and the dynamic importance degrees. The controller can smoothly realize in parallel the angular control of the pendulum and the position control of the cart in a very natural way without reasoning a virtual target angle. For any inverted pendulum and cart systems, of which the pendulum length is between [0.2m, 2.2m], the pendulum mass is larger than or equal to 0.001kg, the cart mass is larger than or equal to 0.002kg, and the mass ratio of the pendulum to the cart is between [0.005, 0.500], the controller is shown to have a very high generalization ability to stabilize the objects completely in about 8 seconds when the initial angle of the pendulum is between [-30.0°, +30.0°], or the initial position of the cart is between [-2.25m, +2.25m].
View full abstract
-
Masatoshi SAKAWA, Ichiro Nishizaki, Yoshio UEMURA
Article type: Article
1999 Volume 11 Issue 3 Pages
503-511
Published: June 15, 1999
Released on J-STAGE: January 08, 2018
JOURNAL
FREE ACCESS
In this paper, we deal with a real problem on production and transportation in a housing material manufacturer, and consider a production and transportation planning under the assumption that the manufacturer makes multiple products at factories in multiple regions and the products are in demand in each of the regions. First, we formulate mixed zero-one programming problems so as to minimize the cost of production and transportation satisfying production capacities and demands. Second, to realize satble production and satisfactory supply of the products, fuzzy programming for the production and transportation problem is incorporated. Finally, under the optimal planning of production and transportation, we show a profit and cost allocation by applying a solution concept from game theory. Practical applications for all the models are given.
View full abstract
-
Tadahiko MURATA, Hisao ISHIBUCHI, Mitsuo GEN
Article type: Article
1999 Volume 11 Issue 3 Pages
512-520
Published: June 15, 1999
Released on J-STAGE: January 08, 2018
JOURNAL
FREE ACCESS
In this paper, first we extend the formulation of multi-objective fuzzy scheduling problems using the ordered weighted averaging (OWA) operator. Two fuzzy scheduling problems have been already proposed: the maximization problem of the total satisfaction grade and the maximization problem of the minimum satisfaction grade. We formulate a multi-objective fuzzy scheduling problem that lies between these two problems. In our multi-objective fuzzy scheduling problem, each job is scheduled according to its own scheduling criteria. Scheduling criteria for each job are denoted by a membership function that indicates a satisfaction grade of a decision maker for the completion time of the corresponding job. Next we introduce job importance grades into the formulation of multi-objective fuzzy scheduling problems. In the previous formulation, all jobs have the same importance. There are, however, many cases where each job has different importance in the scheduling. That is, some jobs may be more important than other jobs. We propose two ways for introducing the importance grades into the formulation of multi-objective fuzzy scheduling problems. Finally we examine solutions(i.e., schedules)obtained in several formulations of multi-objective scheduling problems by computer simulations.
View full abstract
-
Hirohisa AMAN, Torao YANARU, Masahiro NAGAMATSU, Kazunori MIYAMOTO
Article type: Article
1999 Volume 11 Issue 3 Pages
521-527
Published: June 15, 1999
Released on J-STAGE: January 08, 2018
JOURNAL
FREE ACCESS
In this paper, for a class on object oriented programming, we represent its structure using directed graph in which each node corresponds to each member of the class. Then we quantify the dependence relationships among class members, with the weighted closure which is defined here. Using these quantified relationships, through fuzzy graph, we proceed to discuss class structural complexity.
View full abstract
-
1999 Volume 11 Issue 3 Pages
528-531
Published: June 15, 1999
Released on J-STAGE: January 08, 2018
JOURNAL
FREE ACCESS
-
1999 Volume 11 Issue 3 Pages
532-
Published: June 15, 1999
Released on J-STAGE: January 08, 2018
JOURNAL
FREE ACCESS