Learning automaton as a model to simulate and analyse learning behaviour in rats

A technique is presented for using a learning automaton as a model to simulate and analyse learning behaviour in rats, and the usefulness of this model is illustrated. This automaton was proposed as a learning machine by Tsetlin (1961) and refined by Aso and Kimura (1976). Some learning properties of such an automaton are first examined and it is shown that the automaton can be used to simulate a variety of learning behaviour, composed of several kinds of action, by the proper selection of parameters. Secondly, a mathematical model of the learning experiment carried out in our laboratory with rats is formulated as an interactive system between a learning automaton and its environment. Finally, the effectiveness of the mathematical model is discussed and proved through the simulation and analysis of learning behaviour with multiple actions in a specific rat.     

Multilayer polarization elements and their applications to polarimetric studies in vacuum ultraviolet and soft X-ray regions

Multilayer polarization elements and their applications to polarimetric studies in 20～400 eV region are mainly reviewed.General principle of selecting material combinations to get high linear polarizance multilayers of reflection type is given with practical examples,with periodic or non-periodic layer structures depending on the usage.Transmission type is introduced as linear polarizer and phase shifter.Their applications include polarization diagnosis of laboratory optical systems and synchrotron radiation beamlines of linear and circular polarization,magnetic rotation experiments such as Faraday rotation and magnetic Kerr rotation on magnetic films and multilayers,and ellipsometry to measure optical constants of thin films precisely.Polarization analysis of soft X-ray fluorescence using multilayer-coated grating is also mentioned.Finally this review is summarized with outlook of further developments.     

DC Fault Clearing Characteristics of a Bulk Power VSC HVDC Transmission System using DC Circuit Breakers

Development of wind power generation resources is considered in Japan. However, long‐distance transmission lines are necessary to utilize the wind power because most of the wind power resources are distant from load centers. A voltage source converter based high voltage dc transmission (VSC HVDC) system is an option to realize their effective use. A VSC HVDC system equipped with dc circuit breakers has been considered to clear the dc line faults promptly. However, the fault clearing characteristics by the dc circuit breakers have not been analyzed very well for the system with the long distance overhead transmission line. This paper focuses on a fault clearing in the VSC HVDC system with long‐distance overhead line by the dc circuit breaker using a new fault detection method. A 2.4‐GW bipolar VSC HVDC system with 600‐km overhead transmission line is modeled and its operation performance is verified by computer simulations. The simulation results demonstrate the detailed fault clearing characteristics with the long transmission line. Moreover, specifications required for the dc circuit breakers are considered in terms of operation delay, peak current, and absorbed energy.     

Non-linear analysis of nitrogen cycle in aquatic ecosystems

In this paper, s mathematical model of the nitrogen cycle in aquatic ecosystems is proposed for studying eutrophication in bodies of water. It is presented that for some parameter values all equilibrium states of the system become unstable and a stable limit cycle appears. This is explained by a Hopf bifurcation. The conditions under which the Hopf bifurcation occurs are obtained by the detailed analysis of the model. Some results of numerical computations are also presented.     

Feedback transformation of linear systems

In this paper we define a feedback transformation as an operation on a class of linear systems and investigate its properties within the framework of the mathematical theory of general systems. In particular, the class of linear systems produced from an arbitrary fixed linear system by the feedback transformation is characterized. Furthermore, a decoupling condition for linear systems is presented as well as a relation between the class of linear systems that can be decoupled and the class of invertible linear systems.     

Learning algorithms of layered neural networks via extended Kalman filters

Learning algorithms are described for layered feedforward type neural networks, in which a unit generates a real-valued output through a logistic function. The problem of adjusting the weights of internal hidden units can be regarded as a problem of estimating (or identifying) constant parametes with a non-linear observation equation. The present algorithm based on (he extended Kalman filter has just the time-varying learning rate, while the well-known back-propagation (or generalized delta rule) algorithm based on gradient descent has a constant learning rate. From some simulation examples it is shown that when a sufficiently trained network is desired, the learning speed of the proposed algorithm is faster than that of the traditional back-propagation algorithm.     

Optimum dynamic design for beam shapes subjected to a moving concentrated load

We are concerned with an optimum dynamic design of beams subjected to a moving concentrated load with constant speed. The influence of the dynamic behaviour of the beams is considered in a proposed optimum design problem. The optimum shapes of beams are determined by the minimization of two kinds of performance indices. The optimization procedure is performed by non-linear programming on the basis of the exterior penalty function and BFGS methods. Optimization is calculated by the modal coordinates transformation and the numerical integration method     

Structure of the class of linear functional time systems under output feedback transformation

In this paper, we discuss from a qualitative and unified viewpoint the change of system behaviour produced by output feedback primarily to clarify mutual relationships between systems under feedback transformation. We first define the class of all linear functional time systems that are characterized by four system-theoretic properties. Next we define feedback as an operation on a class of systems and discuss the change of system behaviour produced by feedback as a system transformation either between two classes of system or within one class of system. Finally we clarify the structure of a class under feedback transformation in terms of system-theoretic concepts. We obtain three major results which provide a theoretical basis for systems science: (1) we specified the subclass of linear systems on which the feedback transformation can be well defined; (2) we characterize a relation between the system transformation under dynamic output feedback and that under static output feedback; (3) we show that the subclass consisting of all symmetric systems is closed under dynamic output feedback transformation.     

An improvement of a back propagation algorithm by the extended Kalman filter and demand forecasting by layered neural networks

Recently, Watanabe et al. proposed a back propagation algorithm via the extended Kalman filter, in which the learning rate was time-varying. In their algorithm the weights and biases are treated as independent variables. It is, however, natural that the weights and biases are not always independent, and generally have mutual correlation. In this paper, we improve the back propagation algorithm by considering that there is mutual correlation among the weights and bias directly connected to the unit. Through some numerical examples, our improved learning algorithm is compared with Watanabe et al.'s algorithm in learning ability. Furthermore, we consider demand forecasting as a kind of pattern recognition, and propose a demand forecasting method using layered neural networks with the improved learning algorithm. The effectiveness of this demand forecasting method is also discussed through some simulations.     

EFFECTS OF HEAT TREATMENT AND PEPSIN DIGESTION ON TRYPSIN INHIBITORY ACTIVITY AND OVOMUCOID ANTIGENICITY OF JAPANESE QUAIL EGG WHITE

Changes in trypsin inhibitory activity and ovomucoid antigenicity in the egg white of Japanese quail egg were observed after subjecting the quail egg to boiling, saúteing and microwave irradiation. Boiling for 10 min caused about 50% loss of the initial trypsin inhibitory activity. With respect to antigenicity, the boiling caused an initial increase followed by a drop to the original level. Saúteing for 3 min resulted in about 90% loss of the inhibitory activity and about 75% loss of the initial antigenicity. Microwave irradiation for 90 s resulted in about 85% loss of the inhibitory activity and about 80% loss of the initial antigenicity. After pepsin digestion of the egg white boiled for 10 min, 100% of the starting trypsin inhibitory activity was retained even after 24 h digestion, while the antigenicity decreased to about 10% of the initial activity. These results suggest that about 50% of the original trypsin inhibitory and 10% of the original antigenic activity remain after a 24 h pepsin digestion of the egg white boiled for 10 min. Thus, we concluded that intact ovomucoid or ovomucoid degradation products with some trypsin inhibitory activity and antigenicity were introduced into the small intestine after ingestion of the cooked egg white of Japanese quail.