Identification of nonlinear systems with a dynamic recurrent neural network

Two approaches are presented to calculate the weights for a Dynamic Recurrent Neural Network (DRNN) in order to identify the input-output dynamics of a class of nonlinear systems. The number of states of the identified network is constrained to be the same as the number of states of the plant.

Learning in neural networks and stochastic approximation methods with averaging

The problem of adjusting the weights (learning) in multilayer feedforward neural networks (NN) is known to be of a high importance when utilizing NN techniques in various practical applications. The learning procedure is to be performed as fast as possible and in a simple computational fashion, the two requirements which are usually not satisfied practically by the methods developed so far. Moreover, the presence of random inaccuracies are usually not taken into account. In view of these three issues, an alternative stochastic approximation approach discussed in the paper, seems to be very promising.

Adapting the International System of Units to the twenty-first century

We review the proposal of the International Committee for Weights and Measures (Comité International des Poids et Mesures, CIPM), currently being considered by the General Conference on Weights and Measures (Conférences Générales des Poids et Mesures, CGPM), to revise the International System of Units (Le Système International d’Unitès, SI). The proposal includes new definitions for four of the seven base units of the SI, and a new form of words to present the definitions of all the units. The objective of the proposed changes is to adopt definitions referenced to constants of nature, taken in the widest sense, so that the definitions may be based on what are believed to be true invariants. In particular, whereas in the current SI the kilogram, ampere, kelvin and mole are linked to exact numerical values of the mass of the international prototype of the kilogram, the magnetic constant (permeability of vacuum), the triple-point temperature of water and the molar mass of carbon-12, respectively, in the new SI these units are linked to exact numerical values of the Planck constant, the elementary charge, the Boltzmann constant and the Avogadro constant, respectively. The new wording used expresses the definitions in a simple and unambiguous manner without the need for the distinction between base and derived units. The importance of relations among the fundamental constants to the definitions, and the importance of establishing a mise en pratique for the realization of each definition, are also discussed.

Making the least active pay: a simulation of rewards and penalties under Demand Side Participation programs

The orthodox approach for incentivising Demand Side Participation (DSP) programs is that utility losses from capital, installation and planning costs should be recovered under financial incentive mechanisms which aim to ensure that utilities have the right incentives to implement DSP activities. The recent national smart metering roll-out in the UK implies that this approach needs to be reassessed since utilities will recover the capital costs associated with DSP technology through bills. This paper introduces a reward and penalty mechanism focusing on residential users. DSP planning costs are recovered through payments from those consumers who do not react to peak signals. Those consumers who do react are rewarded by paying lower bills. Because real-time incentives to residential consumers tend to fail due to the negligible amounts associated with net gains (and losses) or individual users, in the proposed mechanism the regulator determines benchmarks which are matched against responses to signals and caps the level of rewards/penalties to avoid market distortions. The paper presents an overview of existing financial incentive mechanisms for DSP; introduces the reward/penalty mechanism aimed at fostering DSP under the hypothesis of smart metering roll-out; considers the costs faced by utilities for DSP programs; assesses linear rate effects and value changes; introduces compensatory weights for those consumers who have physical or financial impediments; and shows findings based on simulation runs on three discrete levels of elasticity.

The correlation structure of spatial autoregressions

This paper investigates how the correlations implied by a first-order simultaneous autoregressive (SAR(1)) process are affected by the weights matrix and the autocorrelation parameter. A graph theoretic representation of the covariances in terms of walks connecting the spatial units helps to clarify a number of correlation properties of the processes. In particular, we study some implications of row-standardizing the weights matrix, the dependence of the correlations on graph distance, and the behavior of the correlations at the extremes of the parameter space. Throughout the analysis differences between directed and undirected networks are emphasized. The graph theoretic representation also clarifies why it is difficult to relate properties ofW to correlation properties of SAR(1) models defined on irregular lattices.

Simple solution for a complex problem: Proanthocyanidins, galloyl glucoses and ellagitannins fit on a single calibration curve in high performance-gel permeation chromatography

This study was undertaken to explore gel permeation chromatography (GPC) for estimating molecular weights of proanthocyanidin fractions isolated from sainfoin (Onobrychis viciifolia). The results were compared with data obtained by thiolytic degradation of the same fractions. Polystyrene, polyethylene glycol and polymethyl methacrylate standards were not suitable for estimating the molecular weights of underivatized proanthocyanidins. Therefore, a novel HPLC-GPC method was developed based on two serially connected PolarGel-L columns using DMF that contained 5% water, 1% acetic acid and 0.15 M LiBr at 0.7 ml/min and 50 degrees C. This yielded a single calibration curve for galloyl glucoses (trigalloyl glucose, pentagalloyl glucose), ellagitannins (pedunculagin, vescalagin, punicalagin, oenothein B, gemin A), proanthocyanidins (procyanidin B2, cinnamtannin B1), and several other polyphenols (catechin, epicatechin gallate, epicallocatechin gallate, amentoflavone). These GPC predicted molecular weights represented a considerable advance over previously reported HPLC-GPC methods for underivatized proanthocyanidins. (C) 2011 Elsevier B.V. All rights reserved.

Effects of omega-3 fatty acids on immune function in broiler chickens

There is interest in the enrichment of poultry meat with long-chain n-3 polyunsaturated fatty acids in order to increase the consumption of these fatty acids by humans. However, there is concern that high levels of n-3 polyunsaturated fatty acids may have detrimental effects on immune function in chickens. The effect of feeding increasing levels of fish oil (FO) on immune function was investigated in broiler chickens. Three-week-old broilers were fed 1 of 4 wheat-soybean basal diets that contained 0, 30, 50, or 60 g/kg of FO until slaughter. At slaughter, samples of blood, bursa of Fabricius, spleen, and thymus were collected from each bird. A range of immune parameters, including immune tissue weight, immuno-phenotyping, phagocytosis, and cell proliferation, were assessed. The pattern of fatty acid incorporation reflected the fatty acid composition of the diet. The FO did not affect the weight of the spleen, but it did increase thymus weight when fed at 50 g/kg (P < 0.001). Fish oil also lowered bursal weights when fed at 50 or 60 g/kg (P < 0.001). There was no significant effect of FO on immune cell phenotypes in the spleen, thymus, bursa, or blood. Feeding 60 g/kg of FO significantly decreased the percentage of monocytes engaged in phagocytosis, but it increased their mean fluorescence intensity relative to that of broilers fed 50 g/kg of FO. Lymphocyte proliferation was significantly decreased after feeding broiler chickens diets rich in FO when expressed as division index or proliferation index, although there was no significant effect of FO on the percentage of divided cells. In conclusion, dietary n-3 polyunsaturated fatty acids decrease phagocytosis and lymphocyte proliferation in broiler chickens, highlighting the need for the poultry industry to consider the health status of poultry when poultry meat is being enriched with FO.

Generalized beta generated distributions

This article introduces generalized beta-generated (GBG) distributions. Sub-models include all classical beta-generated, Kumaraswamy-generated and exponentiated distributions. They are maximum entropy distributions under three intuitive conditions, which show that the classical beta generator skewness parameters only control tail entropy and an additional shape parameter is needed to add entropy to the centre of the parent distribution. This parameter controls skewness without necessarily differentiating tail weights. The GBG class also has tractable properties: we present various expansions for moments, generating function and quantiles. The model parameters are estimated by maximum likelihood and the usefulness of the new class is illustrated by means of some real data sets.

Oligosaccharides of pitaya (dragon fruit) flesh and their prebiotic properties

The major carbohydrates of white and red-flesh pitayas (dragon fruit) were glucose, fructose and some oligosaccharides (total concentrations of 86.2 and 89.6 g/kg, respectively). The molecular weight distribution of the extract was affected by the extraction solvent. The maximum oligosaccharides content (27.40%), which included fractions with molecular weights of 273–275, 448–500 and 787–911 Da, were obtained using 80% ethanol extraction at room temperature (28 ± 2 °C). The low molecular weight fraction, including glucose and fructose, was successfully removed by yeast cultivation. The molecular weights of mixed oligosaccharides (716, 700, 490 and 474 Da) were confirmed by mass spectrometry. The mixed oligosaccharides showed that they were resistant to hydrolysis by artificial human gastric juice and human α-amylase, giving maximum hydrolysis of 4.04% and 34.88%, respectively. The mixed oligosaccharides were also found capable of stimulating the growth of lactobacilli and bifidobacteria.

The effect of the molecular weight on the electro-optic properties of methacrylate-based side-chain liquid crystal polymers

A series of methacrylate-based side-chain liquid crystal polymers has been prepared with a range of molecular weights. For the high molecular weight polymers a smectic phase is observed with a very narrow nematic range; however, for low molecular weight polymers only the nematic phase is observed. A marked reduction in the glass transition temperature, TSN and TNI is observed with a reduction in the molecular weight. The orientational order parameters for these polymers in the liquid crystal phase have been determined using infra-red dichroism. It is found that the higher the molecular weight of the polymer, the greater is the threshold voltage of the electro-optic response and the lower the order parameter. The increase in the threshold voltage with increasing molecular weight may be related to the intrinsic curvature elasticity and hence to the coupling between the mesogenic units and the polymer backbone.

A robust controller design method for feedback substitution schemes using genetic algorithms

Controllers for feedback substitution schemes demonstrate a trade-off between noise power gain and normalized response time. Using as an example the design of a controller for a radiometric transduction process subjected to arbitrary noise power gain and robustness constraints, a Pareto-front of optimal controller solutions fulfilling a range of time-domain design objectives can be derived. In this work, we consider designs using a loop shaping design procedure (LSDP). The approach uses linear matrix inequalities to specify a range of objectives and a genetic algorithm (GA) to perform a multi-objective optimization for the controller weights (MOGA). A clonal selection algorithm is used to further provide a directed search of the GA towards the Pareto front. We demonstrate that with the proposed methodology, it is possible to design higher order controllers with superior performance in terms of response time, noise power gain and robustness.

Real estate portfolio construction and estimation risk

The use of MPT in the construction real estate portfolios has two serious limitations when used in an ex-ante framework: (1) the intertemporal instability of the portfolio weights and (2) the sharp deterioration in performance of the optimal portfolios outside the sample period used to estimate asset mean returns. Both problems can be traced to wide fluctuations in sample means Jorion (1985). Thus the use of a procedure that ignores the estimation risk due to the uncertain in mean returns is likely to produce sub-optimal results in subsequent periods. This suggests that the consideration of the issue of estimation risk is crucial in the use of MPT in developing a successful real estate portfolio strategy. Therefore, following Eun & Resnick (1988), this study extends previous ex-ante based studies by evaluating optimal portfolio allocations in subsequent test periods by using methods that have been proposed to reduce the effect of measurement error on optimal portfolio allocations.

Implementation of Parallel (p,q) Counters for High Speed Digital Multipliers

Both the (5,3) counter and (2,2,3) counter multiplication techniques are investigated for the efficiency of their operation speed and the viability of the architectures when implemented in a fast bipolar ECL technology. The implementation of the counters in series-gated ECL and threshold logic are contrasted for speed, noise immunity and complexity, and are critically compared with the fastest practical design of a full-adder. A novel circuit technique to overcome the problems of needing high fan-in input weights in threshold circuits through the use of negative weighted inputs is presented. The authors conclude that a (2,2,3) counter based array multiplier implemented in series-gated ECL should enable a significant increase in speed over conventional full adder based array multipliers.

Modelling and inverting complex-valued Wiener systems

We develop a complex-valued (CV) B-spline neural network approach for efficient identification and inversion of CV Wiener systems. The CV nonlinear static function in the Wiener system is represented using the tensor product of two univariate B-spline neural networks. With the aid of a least squares parameter initialisation, the Gauss-Newton algorithm effectively estimates the model parameters that include the CV linear dynamic model coefficients and B-spline neural network weights. The identification algorithm naturally incorporates the efficient De Boor algorithm with both the B-spline curve and first order derivative recursions. An accurate inverse of the CV Wiener system is then obtained, in which the inverse of the CV nonlinear static function of the Wiener system is calculated efficiently using the Gaussian-Newton algorithm based on the estimated B-spline neural network model, with the aid of the De Boor recursions. The effectiveness of our approach for identification and inversion of CV Wiener systems is demonstrated using the application of digital predistorter design for high power amplifiers with memory

Affinity purification of 61- and 65-kDa rat brain corticotropin-releasing factor receptors and receptor-associated G proteins.

Corticotropin-releasing factor (CRF) has been shown to have a central role in physiological adaptation to stress. It is recognized for stimulating the release of adrenocorticotropin from the anterior pituitary gland, and has more recently been implicated as a regulator of autonomic and immunological responses to stress. Much confusion has surrounded the characterization of CRF receptors, with proteins of varying molecular weights having been identified but never purified and characterized. Recently, two CRF receptors have been cloned from brain and pituitary gland, but evidence from in-situ hybridization studies suggests that further CRF receptor types exist. We therefore developed two techniques which enable the isolation of CRF receptors from whole rat brain. The use of a solid-phase CRF analogue affinity column and elution using a competing ligand resulted in the purification of a single protein of 61 kDa. A second technique was devised which allowed the co-isolation of associated signalling proteins and the identification of CRF bound species following purification. CRF was covalently cross-linked to receptors and the complex purified using antibodies specific for the ligand. This enabled the purification of a CRF receptor of approximately 65 kDa and associated alpha and beta gamma G protein subunits. This study demonstrates the successful isolation of CRF receptors which are of different molecular weights to those previously observed from affinity cross-linking studies or predicted from cloned genes. In addition, we confirm the involvement of G proteins in CRF stimulated cell signalling by demonstrating their association with purified CRF receptor.