Parallel, self-organizing, hierarchical neural networks. II

For pt.I see IEEE Trans. Neural Networks, vol.1, p.167-78 (1990). Parallel, self-organizing, hierarchical neural networks (PSHNNs) involve a number of stages with error detection at the end of each stage, i.e., rejection of error-causing vectors, which are then fed into the next stage after a nonlinear transformation. The stages operate in parallel during testing. Statistical properties and the mechanisms of vector rejection of the PSHNN are discussed in comparison to the maximum likelihood method and the backpropagation network. The PSHNN is highly fault tolerant and robust against errors in the weight values due to the adjustment of the error detection bounds to compensate errors in the weight values. These properties are exploited to develop architectures for programmable implementations in which the programmable parts are reduced to on-off or bipolar switching operations for bulk computations and attenuators for pointwise operations     

Electrochemical Aspects of Electroosmotic Dewatering of Clay Suspensions

An attempt is made to elucidate the electrochemical aspects of the electroosmotic dewatering (EOD) of clays as reported in some recent work, especially that on interrupted DC power electroosmotic dewatering published by Rabie, Mujumdar and Weber (2). These authors showed that the dewatering by EOD stops after the DC power has been on for several minutes or hours; on interruption of their power and on short-circuiting of the electrodes, conditions can be created again for some further dewatering by DC power EOD. This discovery, of Rabie et al. is interpreted as a fuel cell effect and it is shown that it affords clues to several other electrochemical strategies for the possible enhancement of the efficiency of the EOD by DC power.

Further, the open circuit potentials observed by Rabie et al. (2) on the interruption of DC power are given an electrochemical interpretation which leads to quantitative estimates in agreement with the experimental values.     

Erbium-doped silicon and porous silicon for optoelectronics

Silicon forms the backbone of the microelectronics industry, and possibly, of the optoelectronics industry, hitherto dominated by III/V materials. One of the remaining goals is to build an optical source in silicon. Erbium exhibits luminescent 1.54 μm intra-4f transitions in both silicon and porous silicon. This paper reviews the work which has been carried out in this field and discusses some possible additional applications of erbium-doped silicon in optoelectronics, such as a novel on-chip temperature sensor.     

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     

Some factors affecting the electrical properties of carbon-black-loaded butyl rubber during vulcanization

Butyl rubber mixtures loaded with 70 phr general purpose furnace black (GPF) and tetramethyl thiuram disulphide (TMTD)/S as vulcanizing system were prepared. The kinetics of their electrical conductivity development during the vulcanization process were followed by using an especially devised system. It was found that the increase in the electrical conductivity during vulcanization obeys an exponential growth function with time constant τ, which markedly decreases with increasing vulcanization temperature as well as with the efficiency of the vulcanizing system. After completion of the vulcanization process, about 80 min, the samples obtained possess reasonable stability and reproducibility of electrical conductivity.     

Optimum design for buckling of plain and stiffened composite axisymmetric shell panels/shells

The buckling of plain and discretely stiffened composite axisymmetric shell panels/shells made of repeated sublaminate construction is studied using the finite element method. In repeated sublaminate construction, a full laminate is obtained by repeating a basic sublaminate, which has a smaller number of plies. The optimum design for buckling is obtained by determining the layup sequence of the plies in the sublaminate by ranking, so as to achieve maximum buckling load for a specified thickness. For this purpose, a four-noded 48-dof quadrilateral composite thin shell element, together with fully compatible two-noded 16-dof composite meridional and parallel circle stiffener elements are used.     

Numerical investigation of nitriding of refractory metals (with account of the phase constitution diagrams for metal-gas systems)

A mathematical model of dissolution of gas in a metal is suggested with account of phase formation in accordance with the phase constitution diagram (PCD). The stage-by-stage saturation process to the final product formation is shown for an individual particle, through which a reaction wave passes, depending on the diffusion permeability of the metal and solubility conditions that obey Sieverts's law. The effect of the filtration supply of the oxidant to the reaction zone and the process exothermicity on the course of the process is shown.Institute of Structural Macrokinetics, Russian Academy of Sciences, Chernogolovka, Russia. Translated from Inzhenerno-Fizicheskii Zhurnal, Vol. 65, No. 4, pp. 447–450, October, 1993.     

A practical controller for explicit rate congestion control

This paper examines congestion control for explicit rate data networks. The available bit rate (ABR) service category of asynchronous transfer mode (ATM) networks serves as an example system, however, the results of this paper are applicable to other explicit rate systems as well. After a plant model is established, an adaptive control strategy is presented. Several algorithm enhancements are then introduced. These enhancements reduce convergence time, improve queue depth management, and reduce parameter bias. This work differentiates itself from the other contributions in the area of rate-based congestion control in its balanced approach of retaining enough complexity as to afford attractive performance properties, but not so much complexity as to make implementation prohibitively expensive