We present the method for determining the relative importance in AHP (Analytic Hierarchy Process) by using fuzzy numbers fitted to the decision maker's feeling for every modifier. This method can use the fuzzy numbers having membership functions with arbitrary spread to represent the respective fuzzy feeling. We can expect that the results of this method fits the feeling of each decision maker better than the results of existing method.
In this paper, a stimulus-response scheme is proposed in chaotic neural networks with synaptic plasticities and the processes of dynamic learning under external stimuli are investigated. Owing to the refractoriness and the time-hysteresis, fixing abilities of patterns of stimuli become much higher than those based on Hopfield neural network and they also have sensitive dependences on the strength of stimulation. These characteristics turn out to be supported with the chaotic activity by examining the relation between the refractoriness and time-hysteresis, and the (maximum) Lyapunov exponent during the engraving of stimuli on the network. The above results indicate the importance of neural chaos when we try to realize the biologically relevant, real-time learning mechanism interactive with the outside, which is difficult for the Hopfield neural network showing poor response behaviors.
An LTS is a very flexible system that measures the 3D position of a moving target statically, dynamically and contactless, at a high sampling rate and automatically tracks the target in a large working space. However, the orientation measurement by LTS is difficult. Authors developed an LTS which can measure orientation simultaneously with position measurement, but this method needs to keep an interval between the tracking beam from laser interferometer and the reflection beam from the reflector. This LTS can also reduce the influence of the interval quantity over the position measurement during tracking. This paper proposes a 3D position tracking measurement method which can keep the interval and attains the 3D positions with a revision using the reflection beam. The effectivity of the new 3D position measurement method was confirmed by an experiment with the initial distance measurement method proposed in the previous report.
In this paper, we proposed and examined source estimation of EEG (Electroencephalogram) by directed coherence analysis. While ordinary coherence analysis estimates the degree of linear correlation of two time series as a total, directed coherence analysis can extract a correlation with direction in the frequency domain. We estimated information flow between electrodes pasted on a head of subject based on this method. Though the directed coherence analysis is usually applied to analyze the relations between two time series with direction, we proposed an expanded method to estimate the directed coherences between more than two signals at the same time and visualize the signal sources by mapping them. It is confirmed that the estimated positions of signal sources were more adequate than the result of an usual mapping method by simulations and the experimental results.
TRANSYT (A Traffic Network Study Tool) is the most representative and practical simulator for the optimization of such control parameters as cycle length, split and offset in road network. The problem with TRANSYT is that the calculated solution using a hill-climbing method has not necessarily insured the optimum value because it can be found by searching for combinations of distributed variables. Further, it depends on the initial values set as the starting point in searching. This paper describes new approaches for the simulator for optimizing control parameters, in which situations from non congested to oversaturated traffic conditions can be treated. The optimization problem is represented by string model and this formulation enables to introduce such search methods as random search, genetic algorithm and simulated annealing. The experiments which were carried out to verify the usefulness showed that they are superior to the conventional method used in TRANSYT regarding convergency and optimality of solution.
This paper is concerned with the construction of the strict model of the heat conduction phenomena with random inputs. It is well known that the parabolic heat equation has an infinite thermal propagation speed. This fact is a drawback of the parabolic heat conduction model. Since this drawback comes from Fourier's law, by revising Fourier's law from the physical view point, the stochastic hyperbolic heat conduction model with stochastic inputs is proposed. In the stochastic hyperbolic heat conduction model, thermal propagation speed becomes finite. It should be noted that the influence of the input to the hyperbolic heat conduction model is not simply additive, but the term related to the time delay of the thermal propagation appeared. In this paper, the existence of the unique solution of the stochastic hyperbolic heat equation is proved. Finally, the comparison between the solution processes of the parabolic equation and the hyperbolic one is shown through simulation experiments.