A mixture of experts network structure construction algorithm for modelling and control

This paper introduces a new fast, effective and practical model structure construction algorithm for a mixture of experts network system utilising only process data. The algorithm is based on a novel forward constrained regression procedure. Given a full set of the experts as potential model bases, the structure construction algorithm, formed on the forward constrained regression procedure, selects the most significant model base one by one so as to minimise the overall system approximation error at each iteration, while the gate parameters in the mixture of experts network system are accordingly adjusted so as to satisfy the convex constraints required in the derivation of the forward constrained regression procedure. The procedure continues until a proper system model is constructed that utilises some or all of the experts. A pruning algorithm of the consequent mixture of experts network system is also derived to generate an overall parsimonious construction algorithm. Numerical examples are provided to demonstrate the effectiveness of the new algorithms. The mixture of experts network framework can be applied to a wide variety of applications ranging from multiple model controller synthesis to multi-sensor data fusion.

Magnetic discontinuities in the near-earth solar wind: evidence of in-transit turbulence or remnants of coronal structure?

Fluctuations in the solar wind plasma and magnetic field are well described by the sum of two power law distributions. It has been postulated that these distributions are the result of two independent processes: turbulence, which contributes mainly to the smaller fluctuations, and crossing the boundaries of flux tubes of coronal origin, which dominates the larger variations. In this study we explore the correspondence between changes in the magnetic field with changes in other solar wind properties. Changes in density and temperature may result from either turbulence or coronal structures, whereas changes in composition, such as the alpha-to-proton ratio are unlikely to arise from in-transit effects. Observations spanning the entire ACE dataset are compared with a null hypothesis of no correlation between magnetic field discontinuities and changes in other solar wind parameters. Evidence for coronal structuring is weaker than for in-transit turbulence, with only ∼ 25% of large magnetic field discontinuities associated with a significant change in the alpha-to-proton ratio, compared to ∼ 40% for significant density and temperature changes. However, note that a lack of detectable alpha-to-proton signature is not sufficient to discount a structure as having a solar origin.

The structure and evolution of the stratospheric vortex in response to natural forcings

The structure and evolution of the Arctic stratospheric polar vortex is assessed during opposing phases of, primarily, the El Niño–Southern Oscillation (ENSO) and the Quasi-Biennial Oscillation (QBO), but the 11 year solar cycle and winters following large volcanic eruptions are also examined. The analysis is performed by taking 2-D moments of vortex potential vorticity (PV) fields which allow the area and centroid of the vortex to be calculated throughout the ERA-40 reanalysis data set (1958–2002). Composites of these diagnostics for the different phases of the natural forcings are then considered. Statistically significant results are found regarding the structure and evolution of the vortex during, in particular, the ENSO and QBO phases. When compared with the more traditional zonal mean zonal wind diagnostic at 60°N, the moment-based diagnostics are far more robust and contain more information regarding the state of the vortex. The study details, for the first time, a comprehensive sequence of events which map the evolution of the vortex during each of the forcings throughout an extended winter period.

Synthesis and X-ray crystal structure of the copper(I) dicarboxylic acid complex [Cu(η1-bdoaH)(PPh3)3] (bdoaH2 = benzene-1,2- dioxyacetic acid)

The copper(II) complex [Cu(bdoa)(H2O)2] (bdoaH2 = benzene-1,2-dioxyacetic acid) reacts with triphenylphosphine (1:4 mol ratio) to give the colourless copper(I) complex [Cu(η1-bdoaH)(PPh3)3] (1) in good yield. The X-ray crystal structure of the complex shows the copper atom at the centre of a distorted tetrahedron, and is ligated by the phosphorus atoms of the three triphenylphosphines and one carboxylate oxygen atom of the bdoaH− ligand. Significant intermolecular hydrogen-bonding exists between the pendant carboxylate OH function of one molecule and the uncoordinated “ketonic” oxygen of a neighbouring molecule. Complex 1 is non-conducting in chloroform but ionizes readily in acetonitrile. The cyclic voltammogram of an acetonitrile solution of 1 shows a single irreversible anodic peak for the oxidation of the PPh3 ligands and the copper(I) centre, and a single irreversible cathodic peak for the reduction of the bdoaH− ion. IR and mass spectral data for 1 are given.

Fulvalene cyclopentadienyl titanium and zirconium(III) and -(IV) complexes. X-Ray crystal structure of [{Ti(η5-C5H5) Cl}2 (μ-O)(μ-η5-η5-C10H8)]

The reaction of the fulvalene titanium(III) hydride [{Ti(η5-C5H5)(μ-H)}2(μ-η5-η5-C10H8)] (1) with chlorine leads to [{Ti(η5-C5H5)(μ-Cl)}2(μ-η5-η5-C10H8)] (3) and [{Ti(η5-C5H5)Cl2}2(μ-η5-η5-C10H8)] (4). The reaction of 3 with azobenzene, in wet toluene, gives [{Ti(η5-C5H5)Cl}2(μ-O)(μ-η5-η5-C10H8)] (5) and 1,2-diphenyl hydrazine. The alkylation of 4 and the analogous zirconium complex [{Zr(η5-C5H55)Cl2}2(μ-η5-η5-C10H8)] (2) with LiCH2SiMe3 or LiCH3 permits isolation of the tetraalkyl derivatives [{M(η5-C5H5)(CH2SiMe3)2}2(μ-η5-η5-C10H8)] (M  Ti (6); Zr (8)) and [{Ti(η5-C5H5)(CH3)2}2(μ-η5-η5C10H8)] (7). All the new fulvalene compounds were characterized by IR, and 1H and 13C NMR spectroscope, and mass spectra and 5 by X-ray diffraction. The structure of 5 is very similar to that of the comparable TiIV compound [{Ti(η5-C5H5)2Cl}2(μ-O)] except for the smaller TiOTi angle (159.4° against 173.81°) and a significant deviation from linearity.

Complete Experimental Structure Determination of the p(3x2)pg Phase of Glycine on Cu{110}

We present a quantitative low energy electron diffraction (LEED) surface-crystallograpic study of the complete adsorption geometry of glycine adsorbed on Cu{110} in the ordered p(3×2) phase. The glycine molecules form bonds to the surface through the N atoms of the amino group and the two O atoms of the de-protonated carboxylate group, each with separate Cu atoms such that every Cu atom in the first layer is involved in a bond. Laterally, N atoms are nearest to the atop site (displacement 0.41 Å). The O atoms are asymmetrically displaced from the atop site by 0.54 Å and 1.18 Å with two very different O-Cu bond lengths of 1.93 Å and 2.18 Å. The atom positions of the upper-most Cu layers show small relaxations within 0.07 Å of the bulk-truncated surface geometry. The unit cell of the adsorbate layer consists of two glycine molecules, which are related by a glide-line symmetry operation. This study clearly shows that a significant coverage of adsorbate structures without this glide-line symmetry must be rejected, both on the grounds of the energy dependence of the spot intensities (LEED-IV curves) and of systematic absences in the LEED pattern.

Asset allocation, cross-class correlation and the structure of property returns

Practical applications of portfolio optimisation tend to proceed on a “top down” basis where funds are allocated first at asset class level (between, say, bonds, cash, equities and real estate) and then, progressively, at sub-class level (within property to sectors, office, retail, industrial for example). While there are organisational benefits from such an approach, it can potentially lead to sub-optimal allocations when compared to a “global” or “side-by-side” optimisation. This will occur where there are correlations between sub-classes across the asset divide that are masked in aggregation – between, for instance, City offices and the performance of financial services stocks. This paper explores such sub-class linkages using UK monthly stock and property data. Exploratory analysis using clustering procedures and factor analysis suggests that property performance and equity performance are distinctive: there is little persuasive evidence of contemporaneous or lagged sub-class linkages. Formal tests of the equivalence of optimised portfolios using top-down and global approaches failed to demonstrate significant differences, whether or not allocations were constrained. While the results may be a function of measurement of market returns, it is those returns that are used to assess fund performance. Accordingly, the treatment of real estate as a distinct asset class with diversification potential seems justified.

Rapid protein domain assignment from amino acid sequence using predicted secondary structure

The elucidation of the domain content of a given protein sequence in the absence of determined structure or significant sequence homology to known domains is an important problem in structural biology. Here we address how successfully the delineation of continuous domains can be accomplished in the absence of sequence homology using simple baseline methods, an existing prediction algorithm (Domain Guess by Size), and a newly developed method (DomSSEA). The study was undertaken with a view to measuring the usefulness of these prediction methods in terms of their application to fully automatic domain assignment. Thus, the sensitivity of each domain assignment method was measured by calculating the number of correctly assigned top scoring predictions. We have implemented a new continuous domain identification method using the alignment of predicted secondary structures of target sequences against observed secondary structures of chains with known domain boundaries as assigned by Class Architecture Topology Homology (CATH). Taking top predictions only, the success rate of the method in correctly assigning domain number to the representative chain set is 73.3%. The top prediction for domain number and location of domain boundaries was correct for 24% of the multidomain set (±20 residues). These results have been put into context in relation to the results obtained from the other prediction methods assessed

The structure of the optimal income tax in the quasi-linear model

Existing numerical characterizations of the optimal income tax have been based on a limited number of model specifications. As a result, they do not reveal which properties are general. We determine the optimal tax in the quasi-linear model under weaker assumptions than have previously been used; in particular, we remove the assumption of a lower bound on the utility of zero consumption and the need to permit negative labor incomes. A Monte Carlo analysis is then conducted in which economies are selected at random and the optimal tax function constructed. The results show that in a significant proportion of economies the marginal tax rate rises at low skills and falls at high. The average tax rate is equally likely to rise or fall with skill at low skill levels, rises in the majority of cases in the centre of the skill range, and falls at high skills. These results are consistent across all the specifications we test. We then extend the analysis to show that these results also hold for Cobb-Douglas utility.

Heat kernel smoothing via Laplace-Beltrami eigenfunctions and its application to subcortical structure modeling

We present a new subcortical structure shape modeling framework using heat kernel smoothing constructed with the Laplace-Beltrami eigenfunctions. The cotan discretization is used to numerically obtain the eigenfunctions of the Laplace-Beltrami operator along the surface of subcortical structures of the brain. The eigenfunctions are then used to construct the heat kernel and used in smoothing out measurements noise along the surface. The proposed framework is applied in investigating the influence of age (38-79 years) and gender on amygdala and hippocampus shape. We detected a significant age effect on hippocampus in accordance with the previous studies. In addition, we also detected a significant gender effect on amygdala. Since we did not find any such differences in the traditional volumetric methods, our results demonstrate the benefit of the current framework over traditional volumetric methods.

Calibration of structure in a distributed forecasting model for a semiarid flash flood: dynamic surface storage and channel roughness

Flash floods pose a significant danger for life and property. Unfortunately, in arid and semiarid environment the runoff generation shows a complex non-linear behavior with a strong spatial and temporal non-uniformity. As a result, the predictions made by physically-based simulations in semiarid areas are subject to great uncertainty, and a failure in the predictive behavior of existing models is common. Thus better descriptions of physical processes at the watershed scale need to be incorporated into the hydrological model structures. For example, terrain relief has been systematically considered static in flood modelling at the watershed scale. Here, we show that the integrated effect of small distributed relief variations originated through concurrent hydrological processes within a storm event was significant on the watershed scale hydrograph. We model these observations by introducing dynamic formulations of two relief-related parameters at diverse scales: maximum depression storage, and roughness coefficient in channels. In the final (a posteriori) model structure these parameters are allowed to be both time-constant or time-varying. The case under study is a convective storm in a semiarid Mediterranean watershed with ephemeral channels and high agricultural pressures (the Rambla del Albujón watershed; 556 km 2 ), which showed a complex multi-peak response. First, to obtain quasi-sensible simulations in the (a priori) model with time-constant relief-related parameters, a spatially distributed parameterization was strictly required. Second, a generalized likelihood uncertainty estimation (GLUE) inference applied to the improved model structure, and conditioned to observed nested hydrographs, showed that accounting for dynamic relief-related parameters led to improved simulations. The discussion is finally broadened by considering the use of the calibrated model both to analyze the sensitivity of the watershed to storm motion and to attempt the flood forecasting of a stratiform event with highly different behavior.

Structure and magnetic properties of polydisperse ferrofluids: a molecular dynamics study

We study by Langevin molecular dynamics simulations systematically the influence of polydispersity in the particle size, and subsequently in the dipole moment, on the physical properties of ferrofluids. The polydispersity is in a first approximation modeled by a bidisperse system that consists of small and large particles at different ratios of their volume fractions. In the first part of our investigations the total volume fraction of the system is fixed, and the volume fraction phi(L) of the large particles is varied. The initial susceptibility chi and magnetization curve of the systems show a strong dependence on the value of phi(L). With the increase of phi(L), the magnetization M of the system has a much faster increment at weak fields, and thus leads to a larger chi. We performed a cluster analysis that indicates that this is due to the aggregation of the large particles in the systems. The average size of these clusters increases with increasing phi(L). In the second part of our investigations, we fixed the volume fraction of the large particles, and increased the volume fraction phi(S) of the small particles in order to study their influence on the chain formation of the large ones. We found that the average aggregate size formed by large particles decreases when phi(S) is increased, demonstrating a significant effect of the small particles on the structural properties of the system. A topological analysis of the structure reveals that the majority of the small particles remain nonaggregated. Only a small number of them are attracted to the ends of the chains formed by large particles.

On the structure and dynamics of Indian monsoon depressions

ERA-Interim reanalysis data from the past 35 years have been used with a newly-developed feature tracking algorithm to identify Indian monsoon depressions originating in or near the Bay of Bengal. These were then rotated, centralised and combined to give a fully three-dimensional 106-depression composite structure – a considerably larger sample than any previous detailed study on monsoon depressions and their structure. Many known features of depression structure are confirmed, particularly the existence of a maximum to the southwest of the centre in rainfall and other fields, and a westward axial tilt in others. Additionally, the depressions are found to have significant asymmetry due to the presence of the Himalayas; a bimodal mid-tropospheric potential vorticity core; a separation into thermally cold- (~–1.5K) and neutral- (~0K) cores near the surface with distinct properties; and that the centre has very large CAPE and very small CIN. Variability as a function of background state has also been explored, with land/coast/sea, diurnal, ENSO, active/break and Indian Ocean Dipole contrasts considered. Depressions are found to be markedly stronger during the active phase of the monsoon, as well as during La Niña. Depressions on land are shown to be more intense and more tightly constrained to the central axis. A detailed schematic diagram of a vertical cross-section through a composite depression is also presented, showing its inherent asymmetric structure.

Structure of the planetary boundary layer over southeast England: modeling and measurements

The Weather Research and Forecasting model was applied to analyze variations in the planetary boundary layer (PBL) structure over Southeast England including central and suburban London. The parameterizations and predictive skills of two nonlocal mixing PBL schemes, YSU and ACM2, and two local mixing PBL schemes, MYJ and MYNN2, were evaluated over a variety of stability conditions, with model predictions at a 3 km grid spacing. The PBL height predictions, which are critical for scaling turbulence and diffusion in meteorological and air quality models, show significant intra-scheme variance (> 20%), and the reasons are presented. ACM2 diagnoses the PBL height thermodynamically using the bulk Richardson number method, which leads to a good agreement with the lidar data for both unstable and stable conditions. The modeled vertical profiles in the PBL, such as wind speed, turbulent kinetic energy (TKE), and heat flux, exhibit large spreads across the PBL schemes. The TKE predicted by MYJ were found to be too small and show much less diurnal variation as compared with observations over London. MYNN2 produces better TKE predictions at low levels than MYJ, but its turbulent length scale increases with height in the upper part of the strongly convective PBL, where it should decrease. The local PBL schemes considerably underestimate the entrainment heat fluxes for convective cases. The nonlocal PBL schemes exhibit stronger mixing in the mean wind fields under convective conditions than the local PBL schemes and agree better with large-eddy simulation (LES) studies.

Plasma structure within poleward-moving cusp/cleft auroral transients: EISCAT Svalbard radar observations and an explanation in terms of large local time extent of events

We report high-resolution observations of the southward-IMF cusp/cleft ionosphere made on December 16th 1998 by the EISCAT (European incoherent scatter) Svalbard radar (ESR), and compare them with observations of dayside auroral luminosity, as seen at a wavelength of 630 nm by a meridian scanning photometer at Ny Alesund, and of plasma flows, as seen by the CUTLASS (co-operative UK twin location auroral sounding system) Finland HF radar. The optical data reveal a series of poleward-moving transient red-line (630 nm) enhancements, events that have been associated with bursts in the rate of magnetopause reconnection generating new open flux. The combined observations at this time have strong similarities to predictions of the effects of soft electron precipitation modulated by pulsed reconnection, as made by Davis and Lockwood (1996); however, the effects of rapid zonal flow in the ionosphere, caused by the magnetic curvature force on the newly opened field lines, are found to be a significant additional factor. In particular, it is shown how enhanced plasma loss rates induced by the rapid convection can explain two outstanding anomalies of the 630 nm transients, namely how minima in luminosity form between the poleward-moving events and how events can re-brighten as they move poleward. The observations show how cusp/cleft aurora and transient poleward-moving auroral forms appear in the ESR data and the conditions which cause enhanced 630 nm emission in the transients: they are an important first step in enabling the ESR to identify these features away from the winter solstice when supporting auroral observations are not available.