In this paper, we consider a continuous type knapsack problem involving fuzzy random variable coefficients. First, we formulate the problem to maximize a degree of possibility that the objective function value satisfies a fuzzy goal. Since the degree of possibility varies randomly, we consider the model based on the fractile optimization model in stochastic programming. After transforming the problem into the deterministic equivalent problem, we propose a polynomial order algorithm for solving the problem efficiently and obtain the time complexity.
This paper proposes an autonomous agent system that can recognize human intentions. Recently, the computers and the robots (agents) are used at any situation in the human life. However, the users and the systems cannot draw out each ability, because, in the present condition, only some I/O channels are allowed in the interaction between human and those systems. Furthermore, in expansion of the function, the users must remember many techniques. On the other hand, although humans do not use the some languages to communicate, they can still exchange their intentions. They can use various I/O channels combination that will recognize complex intention transmission. Therefore, even if they cannot understand partner's language, they can take some communications. We paid attention to the transmission of information by nonverbal information, such as gesture or body language used generally in human's communications. We applied the method of communications between humans to that between human and systems. And we aimed at organizing the autonomous agent that understands commands and situations by itself and acts, by artificial ontologies that are technological models of human intentions. In this system, each agent constructs artificial ontologies, and shares with other agents. As a result, this system can correspond to the different individual intention expression and environments. Artificial ontologies are constructed by Conceptual Fuzzy Sets, which is based on a bi-directional associative memory network. This system recognizes the human's action by a bi-directional associative memory. We show the effectiveness of the information sharing by artificial ontology in this agent system. The purpose of this research is to create human centered systems that can provide a natural interface in their interaction with people. This paper shows effectiveness to various human support robots.
This research presents a linear solution method for multi-product and multi-period aggregate production planning (APP) with fuzzy goals. Two main objectives of aggregate production planning: the maximization of profit and the minimization of changes of workforce level are defined by fuzzy goals. Fuzzy goals are quantified by concave polyhedral membership functions, which are directly gotten from the decision maker or from linear approximation of continuous concave membership functions. Forecasted demands are also considered to be fuzzy. Trapezoidal membership functions are used to represent them. Linear coordination method, which has been shown its advantages in formulating convex polyhedral penalty functions, is applied to transform a multiple objective APP problem with fuzzy goals and demands to the crisp optimization problem using only linear equations by converse consideration on the maximization problem of concave polyhedral membership functions as the minimization problem of convex polyhedral penalty functions. The efficient linear coordination model is proposed in this research. A satisfactory efficient solution, which is also close to the decision maker's requirements, can be obtained. This model can be easily solved by the existing linear programming solvers. The proposed model is more appropriate than non-fuzzy formulations in terms of reflecting a realistic problem. Moreover, preference information from the decision maker can be clearly exhibited using concave polyhedral membership functions, which have not been used in existing APP problems. Finally, a numerical example is illustrated and briefly discussed.
In general, some equations can be reduced from others (that is, some equations are not independent from others) in a set of axioms of an algebra Then, it is an important problem to find a complete and independent set of axioms, where each axiom is independent of each other. In this paper the research works on axioms of Kleene algebra are surveyed. At first, we show a given axiom of Kleene algebra is independent or not from the others by using the method of indeterminate coefficients. Next, through these investigations we can clarify the properties of each axiom of Kleene algebra and 32 kinds of complete and independent sets of axioms are shown.
Although innumerable research has challenged the design of fuzzy control systems from the standpoint of stability, most of them are just considering the fuzzy rules as an approximator to deal with the unknown functions in the plant to be controlled. Also, the fuzzy rules are changed completely from their origins which are obtained based on the views of experts after the parameters in the fuzzy rules are tuned. Actually, the primary idea of fuzzy control is to employ the knowledge of experts to control a plant instead of the algorithm extracted from the mathematical model of the plant. Clearly, if the original fuzzy rules are changed at all, what is the point of fuzzy control? In this paper, we consider a few cases to show how to develop a stable control system which persistently maintains the fuzzy rules in accordance with experts' views. Also, computer simulations will support the approach appeared in this paper.