Evolving artificial neural network ensembles

Using a coordinated group of simple solvers to tackle a complex problem is not an entirely new idea. Its root could be traced back hundreds of years ago when ancient Chinese suggested a team approach to problem solving. For a long time, engineers have used the divide-and-conquer strategy to decompose a complex problem into simpler sub-problems and then solve them by a group of solvers. However, knowing the best way to divide a complex problem into simpler ones relies heavily on the available domain knowledge. It is often a manual process by an experienced engineer. There have been few automatic divide-and-conquer methods reported in the literature. Fortunately, evolutionary computation provides some of the interesting avenues to automatic divide-and-conquer methods. An in-depth study of such methods reveals that there is a deep underlying connection between evolutionary computation and ANN ensembles. Ideas in one area can be usefully transferred into another in producing effective algorithms. For example, using speciation to create and maintain diversity had inspired the development of negative correlation learning for ANN ensembles, and an in-depth study of diversity in ensembles. This paper will review some of the recent work in evolutionary approaches to designing ANN ensembles.     

A mechanical model for the deformational behavior of the polymeric membranes operating in pressure-driven processes

A mechanical model consisting of three elements (elastic, viscoelastic, and plastic) is proposed to describe the deformational behavior of a polymeric membrane operating in a pressure-driven process. The changes in the membrane thickness are proposed to be calculated from the experimentally obtained flux data. The advantages of a mechanical model over the conventional method of hysteresis area measurement as the mechanical stability of the polymeric membranes are discussed. © 1992 John Wiley & Sons, Inc.     

A pairing-free identity-based authenticated group key agreement protocol for imbalanced mobile networks

The secure and reliable group communication gains popularity in imbalanced mobile networks due to the increase demand of the group-oriented applications such as teleconferences, collaborative workspaces, etc. For acquiring the group security objectives, many authenticated group key agreement (AGKA) protocols exploiting the public key infrastructure have been proposed, which require additional processing and storage space for validation of the public keys and the certificates. In addition, the most of the AGKA protocols are implemented using bilinear pairing and a map-to-point (MTP) hash function. The relative computation cost of the bilinear pairing is approximately two to three times more than the elliptic curve point multiplication (ECPM) and the MTP function has higher computation cost than an ECPM. Due to the limitation of communication bandwidth, computation ability, and storage space of the low-power mobile devices, these protocols are not suitable especially for insecure imbalanced mobile networks. To cope with the aforementioned problems, in this paper, we proposed a pairing-free identity-based authenticated group key agreement protocol using elliptic curve cryptosystem. It is found that the proposed protocol, compared with the related protocols, not only improves the computational efficiencies, but also enhances the security features.     

Impact of high-κ dielectric and metal nanoparticles in simultaneous enhancement of programming speed and retention time of nano-flash memory

A methodology to simulate memory structures with metal nanocrystal islands embedded as floating gate in a high-κ dielectric material for simultaneous enhancement of programming speed and retention time is presented. The computational concept is based on a model for charge transport in nano-scaled structures presented earlier, where quantum mechanical tunneling is defined through the wave impedance that is analogous to the transmission line theory. The effects of substrate-tunnel dielectric conduction band offset and metal work function on the tunneling current that determines the programming speed and retention time is demonstrated. Simulation results confirm that a high-κ dielectric material can increase programming current due to its lower conduction band offset with the substrate and also can be effectively integrated with suitable embedded metal nanocrystals having high work function for efficient data retention. A nano-memory cell designed with silver (Ag) nanocrystals embedded in Al₂O₃ has been compared with similar structure consisting of Si nanocrystals in SiO₂ to validate the concept.     

The growth and characterization of AlGaAs double heterostructures for the evaluation of reactor and source quality

AlGaAs double heterostructures are grown by low-pressure metalorganic chemical vapor deposition to evaluate the level of oxygen contamination in different trimethylaluminum sources. Effects of arsine purifiers, misoriented substrates, atmospheric exposure of the growth chamber, and possible phosphorus contamination are also studied. Extensive characterization is performed on these films by a variety of methods, including high-resolution x-ray diffraction, photoluminescence (PL), time-resolved photoluminescence, and secondary-ion mass spectrometry. The PL intensities for structures grown with the low-alkoxide grade are reproducibly much greater than those grown with the regular-grade TMA1. The use of AsH₃ purification improves the PL intensity.     

Estimation of the Substrate Concentration Profiles in Petroleum Pipelines Containing Biofilms

In industrial systems, the formation of biofilms can cause many problems, such as an increase in the flow resistance of pipelines, energy losses in fluid transport and heat exchangers, product contamination, materials deterioration, and biocorrosion. As a result, biofilms contribute substantially to economic losses in the industry. Corrosion is particularly an issue in the petroleum industry and its implications range from down-hole completion through petroleum processing units. Much of this corrosion is attributed to microbial activities. This paper proposes a mathematical model for predicting substrate concentration for such microbial growth. Substrate concentrations in the system and near the biofilm surface are one of the parameters that has a great effect in determining the extent of the problems associated with biofilms. In this study, a convective-diffusion model under various flow conditions (stagnant, laminar, and turbulent) has been solved using the finite difference technique, employing the alternating direction implicit (ADI) method. The model assumes that a liquid containing substrate and bacteria is flowing in a pipeline with known concentrations at the inlet and then predicts the variation of the transient (as a function of time) substrate concentration along the pipeline and as a function of the pipe radius. The model is then used to predict and estimate the substrate concentration profiles on the biofilm surface under different environmental conditions. A parametric study was also conducted to study the effect of the different parameters influencing the substrate concentration profiles in the system and on the biofilm surface.     

Novel Near‐Infrared Squaraine Sensitizers for Stable and Efficient Dye‐Sensitized Solar Cells

A new molecular design strategy for tuning the energy levels of cis‐configured squaraine sensitizers for dye‐sensitized solar cells is described. The Hammett substituent constant and the π‐conjugation length are used as quantitative indicators to modify the central squarate moiety of the sensitizer dyes; specifically, novel near‐infrared squaraine dyes HSQ3 and HSQ4 are synthesized by incorporation of an electron‐withdrawing and π‐extending ethyl cyanoacetate unit on the central squarate moiety. The solution absorption maximum of HSQ4 occurs at 703 nm, and the energy levels of the lowest unoccupied molecular orbital and the highest occupied molecular orbital are in the ideal range for energetically efficient electron injection and regeneration of the oxidized dye. A solar cell sensitized with HSQ4 exhibits a broad incident photo­n‐to‐current conversion efficiency spectrum, extending into the near‐infrared region with a maximum value of 80% at 720 nm, which is is the highest value reported for a squaraine dye–based dye‐sensitized solar cell. The HSQ4‐sensitized solar cell also exhibits excellent durability during light soaking, owing to the double anchors attaching the dye to the TiO₂ surface and to the long alkyl chains extending outward from the surface.     

Design and Analysis of InGaN-GaN Modulation Doped Field-Effect Transistors (MODFETs) for Over 60 GHz Operation

A Modulation-Doped Field-Effect Transistor (MODFET) structure realized in InGaN-GaN material system is presented for the first time. An analytical model predicting the transport characteristics of the proposed MODFET structure is given in detail. Electron energy levels inside and outside the quantum well channel of the MODFET are evaluated. The two-dimensional electron gas (2DEG) density in the channel is calculated by self-consistently solving Schrödinger and Poisson's equations simultaneously. Analytical results of the current-voltage and transconductance characteristics are presented. The unity-current gain cutoff frequency (f _T) of the proposed device is computed as a function of the gate voltage V _G . The results are compared well with experimental f _T value of a GaN/AlGaN HFET device. By scaling the gate length down to 0.25 μm the proposed InGaN-GaN MODFET can be operated up to about 80GHz. It is shown in this paper that InGaN-GaN system has small degradation in f _T as the operating temperature is increased from 300°K to 400°K.     

Novel Engineered Ion Channel Provides Controllable Ion Permeability for Polyelectrolyte Microcapsules Coated with a Lipid Membrane

The development of nanostructured microcapsules based on a biomimetic lipid bilayer membrane (BLM) coating of poly(sodium styrenesulfonate) (PSS)/poly(allylamine hydrochloride) (PAH) polyelectrolyte hollow microcapsules is reported. A novel engineered ion channel, gramicidin (bis‐gA), incorporated into the lipid membrane coating provides a functional capability to control transport across the microcapsule wall. The microcapsules provide transport and permeation for drug‐analog neutral species, as well as positively and negatively charged ionic species. This controlled transport can be tuned for selective release biomimetically by controlling the gating of incorporated bis‐gA ion channels. This system provides a platform for the creation of “smart” biomimetic delivery vessels for the effective and selective therapeutic delivery and targeting of drugs.