This paper discusses some problems of probabilistic methods for system reliability analysis from a fuzzy set theoretical point of view. Probabilistic methods have been developed to analyze system reliability in an objective manner ; it is said that deterministic methods, which have been used before the invention of probabilistic methods, are heavily affected by subjective views of problem analysts. This paper shows that the problem has not yet been solved completely even by the probabilistic methods. Next this paper reviews some applications of fuzzy set theory to system reliability analysis. Especially, it discusses fuzziness of system state, fuzzy probability, utilization of fuzzy set theory in linguistic approach to system reliability analysis, applications of fuzzy measures and fuzzy integrals to modeling of relations between performance shaping factors and reliability.

View full abstract

Organizational accidents have occurred repeatedly in Japan over the past several years. Most of these accidents seem to have been occurred within a gray zone between the safety zone and the danger zone rather than within the danger zone. The present paper considers "risk engineering" as engineering for preventing such accidents that occur within the gray zone, and proposes a framework for analyzing design methods of reducing the gray zone. Firstly, it asserts that safety management design should be classified into a safety-assurance design, and a danger-avoidance design and, that the danger-avoidance design method should incorporate two types of learning mechanism. Next, our paper shows two positive examples utilizing a gray zone for reducing risk ; one is a description method for operation manual or rules, and the other is a reliable safety monitoring design method.

View full abstract

A class of polymodal logic is considered : a Kripke model and an axiomatic system are described and a theorem of completeness is mentioned. Application to a class of systems with uncertainty including risky situations is discussed. Correspondence of components of polymodal logic to the system elements is described. Simple illustrative examples are shown.

View full abstract

[in Japanese]

View full abstract

[in Japanese]

View full abstract

This paper deals with behavioral evolution of multiple robots in a quasi ecosystem. An ecosystem model composed of insects and plants, which is in a relationship of parasitism, is simulated on a cell space. In this ecosystem, the plants become easy to be eliminated as the population size of the insects increases. Consequently, it is required to maintain numerical balance of plants and insects in the quasi-ecosystem. Therefore, multiple robots are introduced to remove insects suitable to the facing environmental condition in the quasi ecosystem. In this paper, we use simple if-then rules and apply genetics-based machine learning for acquiring a strategy for removing insects. Furthermore, we show several simulation results of behavior learning of multiple robots.

View full abstract

In this paper, a fuzzy controller with the capability of compensating the influence of faults is discussed. A stuck-at fault in the membership function is assumed to be fault models. The proposed inference scheme locates a candidate regarded as faulty function, and then exchanges its probably false degree for one of the following values : O, the degree in the next function to the candidate, and the difference between the constant and the degree in the next function to the candidate. If a fault occurs in the consequent part, the proposed scheme shifts several fuzzy variables, and then forms the inference result by using the membership functions allocated newly to the variables. The modified deterministic output of the controller is obtained by shifting the center of gravity of the inference result. The amounts of these shifts are determined systematically. Experimental results for a commercial controller show that the proposed scheme is valid for any single stuck-at fault deviating the normal deterministic output of the controller.

View full abstract

In this paper, in case which it turns around to distribute the supplies to multiple customers of corporate activity, we describe how to specify an efficient transportation route and how to calculate the transportation time-schedule of supplies on it and the anticipation time of arrival to each customer. Now, in fact, the travel times between customers change at any time in same traveling route, so we allow minimization of the total round time, and also consider the specification of arrival time to a customer and the restrictions conditions of transportation order. By these conditions, a transportation route is determined by evaluating the fuzzy value indices for the total round time and the arrival times, which are defined by fuzzy numbers. In this decision, it is employed a genetic algorithm to search the best transportation route and to reduce the searching time. Furthermore, the information of the mesh map are used for calculations of moving-times of the transportations, that is, on the basis of the moving time between the mesh zones and the moving time in the mesh area, an extend-method of Dijkstra is employed to calculate the moving times which changes at every moment. Therefore, we introduced a genetic algorithm and a fuzzy theory to this scheduling, and also describe how to display the best route obtained by that. Finally, from the simulation results of the fuzzy scheduling, we discuss the usefulness of this method.

View full abstract

There are two kinds of games, a game with perfect information and a game with imperfect information. In the study on a game with perfect information such as Shogi, Chess, strong computer programs of Shogi, Chess have been developed, which win a world champion or are as strong as amateurs with grade holders. On the other hand in the study on a game with imperfect information such as poker, contract bridge, strong algorithms are not developed yet since assessment of the strength of a hand and opponent strategy using imperfect information has much uncertainty. This paper takes a Seven card Stud Poker game as a game with imperfect information and aims at the construction of its playing system. The present system has fuzzy rules with non-fixed consequent parts and with threshold values in order to make decisions according to opponent strategy. The threshold values are changed according to opponent strategy during the game. The system is a decision making system that assesses superiority/inferiority of its own hand, and makes decisions whether it drops or not and how much it bets in a game by the fuzzy inference using these irregular fuzzy rules. The usefulness of the system is discussed by analyzing data obtained in experiments, in which the present system plays a stud poker game with human players.

View full abstract