Second Order Optimality Conditions in Nonsmooth Unconstrained Optimization

AMS subject classification: 49J52, 90C30.

Well-Behavior, Well-Posedness and Nonsmooth Analysis

AMS subject classification: 90C30, 90C33.

An Iterative Procedure for Solving Nonsmooth Generalized Equation

2000 Mathematics Subject Classification: 47H04, 65K10.

Topology Studies of Hydrodynamics Using Two-Particle Correlation Analysis

The effects of fluctuating initial conditions are studied in the context of relativistic heavy ion collisions where a rapidly evolving system is formed. Two-particle correlation analysis is applied to events generated with the NEXSPHERIO hydrodynamic code, starting with fluctuating nonsmooth initial conditions (IC). The results show that the nonsmoothness in the IC survives the hydroevolution and can be seen as topological features of the angular correlation function of the particles emerging from the evolving system. A long range correlation is observed in the longitudinal direction and in the azimuthal direction a double peak structure is observed in the opposite direction to the trigger particle. This analysis provides clear evidence that these are signatures of the combined effect of tubular structures present in the IC and the proceeding collective dynamics of the hot and dense medium.

Functionalisation of nanoparticles as electrolytes in fuel cells

Inorganic metal oxide materials are generally poor proton conductors as conductivities are lower than 10-5-10-6 S.cm-1. However, by functionalising Silica, Zirconia or Titania, proton conduction increases by up to 5 orders of magnitude. Hence, functionalised nanomaterials are becoming very competitive against conventional electrolyte materials such as Nafion. In this work, sol-gel processes are employed to produce silica phosphate, zirconia phosphate and titania phosphate functionalised nanoparticles. Furthermore, conductivities at hydrate conditions are investigated, and nanoparticle formation and functionalisation effects on proton conductivity are discussed. Results show conductivities up to 10-1 S.cm-1 (95% RH). Proton conduction increases with the functionalisation content, however heat treatment of nanoparticles locks the functionality in the crystal phase, thus inhibiting proton conduction. Controlling the mesopore phase allows for high proton conduction at hydrated conditions, clearly indicating facilitated ion transport through the pore channels.

Thermal effects in the evolution of initially layered mantle material

A simplified model for anisotropic mantle convection based on a novel class of rheologies, originally developed for folding instabilities in multilayered rock (MUHLHAUS et al., 2002), is extended ¨ through the introduction of a thermal anisotropy dependent on the local layering. To examine the effect of the thermal anisotropy on the evolution of mantle material, a parallel implementation of this model was undertaken using the Escript modelling toolkit and the Finley finite-element computational kernel (DAVIES et al., 2004). For the cases studied, there appears too little if any effect. For comparative purposes, the effects of anisotropic shear viscosity and the introduced thermal anisotropy are also presented. These results contribute to the characterization of viscous anisotropic mantle convection subject to variation in thermal conductivities and shear viscosities.

Growth Response of Various Perennial Grasses to Increasing Salinity

Although the effect of salinity on plant growth has been the focus of a substantive research effort, much of this research has failed to adequately separate the various growth limiting aspects of salinity; thus the results are confounded by multiple factors. Eight perennial grass species were grown in a sand culture system dominated by NaCl (electrical conductivities (ECs) between 1.4 and 38 dS m 1), with sufficient Ca added to each treatment to ensure that Na-induced Ca deficiency did not reduce growth. Of the eight perennial grass species examined, Chloris gayana cv. Pioneer (Rhodes grass) was the most salt tolerant species, whilst in comparison, Chrysopogon zizanioides cv. Monto (vetiver) was of only moderate tolerance. However, observed salinity tolerances tended to be lower than those expected from published values based on the threshold salinity model (bent stick model). This discrepancy may be due in part to differences in the evapotranspirational demand between studies; an increase in demand accelerating the accumulation of Na in the shoots and hence decreasing apparent salinity tolerance. It was also observed that the use of a non-saline growth period to allow seed germination and establishment results in the overestimation of vegetative salinity tolerance if not taken into consideration. This is particularly true for species of low salt tolerance due to their comparatively rapid growth in the non-saline medium compared to that at full salinity.

Wall material properties of yeast cells. Part II. Analysis

In the preceding paper (Part I) force-deformation data were measured with the compression experiment in conjunction with the initial radial stretch ratio and the initial wall-thickness to cell-radius ratio for baker's yeast (Saccharomyces cerevisiae). In this paper, these data have been analysed with the mechanical model of Smith et al. (Smith, Moxham & Middelberg (1998) Chemical Engineering Science, 53, 3913-3922) with the wall constitutive behaviour defined a priori as incompressible and linear-elastic. This analysis determined the mean Young's modulus ((E) over bar), mean maximum von Mises stress-at-failure (<(sigma)over bar>(VM,f)) and mean maximum von Mises strain-at failure (<(epsilon)over bar>(VM,f)) to be (E) over bar = 150 +/- 15 MPa, <(sigma)over bar>(VM,f) = 70 +/- 4 MPa and <(epsilon)over bar>(VM,f) = 0.75 +/- 0.08, respectively. The mean Young's modulus was not dependent (P greater than or equal to 0.05) on external osmotic pressure (0-0.8 MPa) nor compression rate (1.03-7.68 mu m/s) suggesting the incompressible linear-elastic relationship is representative of the actual cell-wall constitutive behaviour. Hydraulic conductivities were also determined and were comparable to other similar cell types (0-2.5 mu m/MPa s). The hydraulic conductivity distribution was not dependent on external osmotic pressure (0-0.8 MPa) nor compression rate (1.03-7.68 mu m/s) suggesting inclusion of cell-wall permeability in the mechanical model is justified. <(epsilon)over bar>(VM,f) was independent of cell diameter and to a first-approximation unaffected (P greater than or equal to 0.01) by external osmotic pressure and compression rate, thus providing a reasonable failure criterion. This criterion states that the cell-wall material will break when the strain exceeds <(epsilon)over bar>(VM,f) = 0.75 +/- 0.08. Variability in overall cell strength during compression was shown to be primarily due to biological variability in the maximum von Mises strain-at-failure. These data represent the first estimates of cell-wall material properties for yeast and the first fundamental analysis of cell-compression data. They are essential for describing cell-disruption at the fundamental level of fluid-cell interactions in general bioprocesses. They also provide valuable new measurements for yeast-cell physiologists. (C) 2000 Elsevier Science Ltd. All rights reserved.

Strain-dependent fluid flow defined through rock mass classification schemes

Strain-dependent hydraulic conductivities are uniquely defined by an environmental factor, representing applied normal and shear strains, combined with intrinsic material parameters representing mass and component deformation moduli, initial conductivities, and mass structure. The components representing mass moduli and structure are defined in terms of RQD (rock quality designation) and RMR (rock mass rating) to represent the response of a whole spectrum of rock masses, varying from highly fractured (crushed) rock to intact rock. These two empirical parameters determine the hydraulic response of a fractured medium to the induced-deformations The constitutive relations are verified against available published data and applied to study one-dimensional, strain-dependent fluid flow. Analytical results indicate that both normal and shear strains exert a significant influence on the processes of fluid flow and that the magnitude of this influence is regulated by the values of RQD and RMR.

Thermal Characteristics of the Mud Nests of the Social Wasp Polybia spinifex (Hymenoptera; Vespidae)

The thermal characteristics of mud nests of Polybia spinifex were investigated by measuring internal and surface temperatures of the nests. The nests had a layer of mud envelope and consisted of mud with fine sand particles. The envelope had a vertically long slit-like entrance hole. The mud nests had high thermal conductivities (0.51-0.67 W/(m degrees C)) comparable to brick, rather than insulation materials of wasps` nests such as paper and wood. It was revealed that the long entrance radiated more heat from the thereto-image. The rate of thermal radiation (emissivity) of the nest surface was 0.80, and the value was similar to that of sand. The internal temperatures of the nests were high from top (T(n1), temperature difference between ambient temperature (T(a)) was 10 degrees C) to bottom (T(n3), difference, 5 degrees C) at noon. On the other hand, the temperature distributions were reversed during the night. Temperature T(n1) was lower by 1 degrees C than T(a), possibly from nightly dew on the top surface, whereas, at the middle point (T(n2),) and T(n3), temperatures were higher by 1 degrees C compared to T(a). Temperature fluctuations (ranges between maximum and minimum temperature) at T(n2) and T(n3) were similar to that of T(a), whereas the values at T(n1) and T(s) were higher than that of T(a).

The role of disorder on the AC and DC electrical conductivity of vapour grown carbon nanofibre/epoxy composites

Four dispersion methods were used for the preparation of vapour grown carbon nanofibre (VGCNF)/epoxy composites. It is shown that each method induces certain levels of VGCNF dispersion and distribution within the matrix, and that these have a strong influence on the composite electrical properties. A homogenous VGCNF dispersion does not necessarily imply higher electrical conductivity. In fact, it is concluded that the presence of well distributed clusters, rather than a fine dispersion, is more important for achieving larger conductivities for a given VGCNF concentration. It is also found that the conductivity can be described by a weak disorder regime.

Hydraulic conductivity in response to exchangeable sodium percentage and solution salt concentration

Hydraulic conductivity is determined in laboratory assays to estimate the flow of water in saturated soils. However, the results of this analysis, when using distilled or deionized water, may not correspond to field conditions in soils with high concentrations of soluble salts. This study therefore set out to determine the hydraulic conductivity in laboratory conditions using solutions of different electrical conductivities in six soils representative of the State of Pernambuco, with the exchangeable sodium percentage adjusted in the range of 5-30%. The results showed an increase in hydraulic conductivity with both decreasing exchangeable sodium percentage and increasing electrical conductivity in the solution. The response to the treatments was more pronounced in soils with higher proportion of more active clays. Determination of hydraulic conductivity in laboratory is routinely performed with deionized or distilled water. However, in salt affected soils, these determinations should be carried out using solutions of electrical conductivity different from 0 dS m-1, with values close to those determined in the saturation extracts.

Direct multisearch for multiobjective optimization

In practical applications of optimization it is common to have several conflicting objective functions to optimize. Frequently, these functions are subject to noise or can be of black-box type, preventing the use of derivative-based techniques. We propose a novel multiobjective derivative-free methodology, calling it direct multisearch (DMS), which does not aggregate any of the objective functions. Our framework is inspired by the search/poll paradigm of direct-search methods of directional type and uses the concept of Pareto dominance to maintain a list of nondominated points (from which the new iterates or poll centers are chosen). The aim of our method is to generate as many points in the Pareto front as possible from the polling procedure itself, while keeping the whole framework general enough to accommodate other disseminating strategies, in particular, when using the (here also) optional search step. DMS generalizes to multiobjective optimization (MOO) all direct-search methods of directional type. We prove under the common assumptions used in direct search for single objective optimization that at least one limit point of the sequence of iterates generated by DMS lies in (a stationary form of) the Pareto front. However, extensive computational experience has shown that our methodology has an impressive capability of generating the whole Pareto front, even without using a search step. Two by-products of this paper are (i) the development of a collection of test problems for MOO and (ii) the extension of performance and data profiles to MOO, allowing a comparison of several solvers on a large set of test problems, in terms of their efficiency and robustness to determine Pareto fronts.

GLODS: Global and Local Optimization using Direct Search

Locating and identifying points as global minimizers is, in general, a hard and time-consuming task. Difficulties increase in the impossibility of using the derivatives of the functions defining the problem. In this work, we propose a new class of methods suited for global derivative-free constrained optimization. Using direct search of directional type, the algorithm alternates between a search step, where potentially good regions are located, and a poll step where the previously located promising regions are explored. This exploitation is made through the launching of several instances of directional direct searches, one in each of the regions of interest. Differently from a simple multistart strategy, direct searches will merge when sufficiently close. The goal is to end with as many direct searches as the number of local minimizers, which would easily allow locating the global extreme value. We describe the algorithmic structure considered, present the corresponding convergence analysis and report numerical results, showing that the proposed method is competitive with currently commonly used global derivative-free optimization solvers.