Evaluation of multicast routing algorithms for real-timecommunication on high-speed networks

Multicast (MC) routing algorithms capable of satisfying the quality of service (QoS) requirements of real-time applications will be essential for future high-speed networks. We compare the performance of all of the important MC routing algorithms when applied to networks with asymmetric link loads. Each algorithm is judged based on the quality of the MC trees it generates and its efficiency in managing the network resources. Simulation results over random networks show that unconstrained algorithms are not capable of fulfilling the QoS requirements of real-time applications in wide-area networks. Simulations also reveal that one of the unconstrained algorithms, reverse path multicasting (RPM), is quite inefficient when applied to asymmetric networks. We study how combining routing with resource reservation and admission control improves the RPM's efficiency in managing the network resources. The performance of one semiconstrained heuristic, MSC, three constrained Steiner tree (CST) heuristics, Kompella, Pasquale, and Polyzos (1992), constrained adaptive ordering (CAO), and bounded shortest multicast algorithm (BSMA), and one constrained shortest path tree (CSPT) heuristic, the constrained Dijkstra heuristic (CDKS) are also studied. Simulations show that the semiconstrained and constrained heuristics are capable of successfully constructing MC trees which satisfy the QoS requirements of real-time traffic. However, the cost performance of the heuristics varies. The BSMA's MC trees are lower in cost than all other constrained heuristics. Finally, we compare the execution times of all algorithms, unconstrained, semiconstrained, and constrained     

Demographic study of first-ever stroke and acute myocardial infarction in Okinawa, Japan

We performed a cross-sectional survey of stroke and acute myocardial infarction (AMI) in Okinawa, Japan with a census population of about 1.2 million. A total of 3,644 cases of first-ever stroke and 898 cases of initial acute myocardial infarction (AMI) were detected. The age-adjusted annual incidence rate for stroke was 105 per 100,000 standard population of Japan, and that of AMI was 26. The case-fatality rate of stroke within 28 days of onset was 12.8%, and that of AMI was 22.2%. Of the stroke cases, 51.4% were diagnosed as brain infarction, 38.7% as brain hemorrhage, and 9.3% as subarachnoid hemorrhage. The diagnosis of stroke subtypes were confirmed by computed tomography or magnetic resonance imaging in 98.4% of all stroke cases. In Okinawa, the incidence rate of AMI was still considerably lower than that in the Western population, and the rate of stroke was similar to that in the Western population.     

Transcutaneous measurement and spectrum analysis of heart wallvibrations

For the noninvasive diagnosis of heart disease based on the acoustic and elastic characteristics of the heart muscle, it is necessary to transcutaneously measure small vibration signals, including components with an amplitude of less than 100 μm, from various parts of the heart wall continuously for periods of more than several heartbeats in a wide frequency range up to 1 kHz. Such measurement, however, has not been realized by any ultrasonic diagnostic methods or systems to date. By introducing the constraint least-square approach, this paper proposes a new method for accurately tracking the movement of the heart wall based on both the phase and magnitude of the demodulated signal to determine the instantaneous position of the object so that the vibration velocity of the moving object can be accurately estimated. By this method, small vibrations of the heart wall with small amplitudes less than 100 μm on the motion resulting from a heartbeat with large amplitude of 10 mm can be successfully detected with sufficient reproducibility in the frequency range up to several hundred Hertz continuously for periods of about 10 heartbeats. The resultant small vibration is analyzed not only in the time domain, but also in the frequency domain. As confirmed by the preliminary experiments herein reported, the new method offers potential for research in acoustical diagnosis of heart disease     

Fuzzy inference and fuzzy inference processor

Fuzzy inference, a data processing method based on the fuzzy theory that has found wide use in the control field, is reviewed. Consumer electronics, which accounts for most current applications of this concept, does not require very high speeds. Although software running on a conventional microprocessor can perform these inferences, high-speed control applications require much greater speeds. A fuzzy inference date processor that operates at 200000 fuzzy logic inferences per second and features 12-b input and 16-b output resolution is described     

Oxygen absorption kinetics in YBa2Cu3Ox

The oxygen absorption process in highly aligned YBa₂Cu₃O_x samples in air and in pure oxygen was studied at several constant temperatures. The process can be expressed in two steps, a chemical reaction controlled step and a diffusion controlled step. The kinetic equations are as follows respectively:

(1 − a_t) ln(1 − a_t) + a_t = k_Dt

The activation energy was calculated using an Arrhenius equation. The technical conditions of heat treatment for oxygen absorption of highly aligned YBa₂Cu₃O_x crystals are discussed.     

Measurements of distribution of self-bias potential on an RF-plane electrode in plasma etching devices

The self-bias potential (V_dc) induced on an RF-powdered electrode (153 mm Ø) in a plasma is measured using electrical probes which are buried in, de-insulated from, and RF-connected to the electrode. The configuration of the probes allows to study the distribution of V_dc discretely on the electrode. The potential is homogeneous in the absence of external magnetic field. In the presence of a homogeneous magnetic field parallel to the electrode, it is reduced and a monotonous gradient takes place in its distribution due to the plasma shift induced by E × B drift. When the magnetic field is rotated along the axis of the RF-electrode at a frequency less than 50 Hz, the distribution, which is almost identical to the one in a static field, rotates with the magnetic field. On the coordinate system rotating with the magnetic field, the probes are regarded to be rotating. The potential distribution is obtained as a continuous function of the azimuthal angle. Thus the rotation of the field provides information for the experimental interpolation.     

Novel technique for SiO2 formed by liquid-phasedeposition for low-temperature processed polysilicon TFT

A technique for SiO₂ formation by liquid-phase deposition (LPD) at nearly room temperature for low-temperature processed (LTP) polysilicon thin-film transistor (poly-Si TFT) was developed. LPD SiO₂ film with a lower P-etch rate shows a dense structure. LPD SiO₂ also exhibits good electrical characteristics. LTP poly-Si thin-film transistors (TFTs) with LPD SiO ₂ as the gate insulator have been fabricated and investigated. Their characteristics indicate performance adequate for their use as pixel transistors in liquid crystal displays (LCDs)     

The importance of electrochemistry-related etching in thegate-groove fabrication for InAlAs/InGaAs HFETs

For the first time, the surface metal on nonalloyed ohmic electrodes is found to significantly change the profiles of gate grooves, when resist openings are employed to monitor drain current during wet-chemical gate recess for sub-micron InAlAs/lnGaAs heterojunction field-effect transistors (HFETs). The surface metal of Ni enhances the etching rate in comparison with that in the absence of electrodes by a factor of 4 and 10, laterally and vertically, which is favorable to fabricate deep gate grooves with small side etching. The Pt surface metal, however, leads to preferential etching of InGaAs over InAlAs, which can be useful to realize large side etching. The existence of an electrochemistry-related etching component, which arises when the ohmic electrodes are present during recess etching, is considered to be responsible for these behaviors     

Learning long-term dependencies with gradient descent is difficult

Recurrent neural networks can be used to map input sequences to output sequences, such as for recognition, production or prediction problems. However, practical difficulties have been reported in training recurrent neural networks to perform tasks in which the temporal contingencies present in the input/output sequences span long intervals. We show why gradient based learning algorithms face an increasingly difficult problem as the duration of the dependencies to be captured increases. These results expose a trade-off between efficient learning by gradient descent and latching on information for long periods. Based on an understanding of this problem, alternatives to standard gradient descent are considered.