Alternating local search based VNS for linear classification

We consider the linear classification method consisting of separating two sets of points in d-space by a hyperplane. We wish to determine the hyperplane which minimises the sum of distances from all misclassified points to the hyperplane. To this end two local descent methods are developed, one grid-based and one optimisation-theory based, and are embedded into a VNS metaheuristic scheme. Computational results show these approaches to be complementary, leading to a single hybrid VNS strategy which combines both approaches to exploit the strong points of each. Extensive computational tests show that the resulting method can always be expected to approach the global optimum close enough that any deviations from the global optimum are irrelevant with respect to the classification power.     

Sauna,sweat and science II – do we sweat what we drink?

ABSTRACT

Inspired by a previous ‘Sauna, sweat and science’ study [Zech et al. Isot Environ Health Stud. 2015;51(3):439–447] and out of curiosity and enthusiasm for stable isotope and sauna research we aimed at answering the question ‘do we sweat (isotopically) what we drink’? We, therefore, pulse-labelled five test persons in a sauna experiment with beverages that were ²H-enriched at about +25,600?‰. Sweat samples were collected during six sauna rounds and the hydrogen isotope composition δ²H_sweat was determined using an isotope ratio mass spectrometer. Before pulse labelling, δ²H_sweat – reflecting by approximation body water – ranged from –32 to –22?‰. This is ～35?‰ enriched compared to usual mid-European drinking water and can be explained with hydrogen-bearing food as well as with the respiratory loss of ²H-depleted vapour. The absence of a clearly detectable ²H pulse in sweat after pulse labelling and δ²H_sweat results of ≤+250?‰ due to a fast ²H equilibration with body water are moreover a clearly negative answer to our research question also in a short-term consideration. Given that the recovery of the tracer based on an isotope mass balance calculation is clearly below 100?%, we finally answer the question ‘where did the rest of the tracer go?’     

Directly testing the linearity assumption for assay validation

The ICH Q2(R1) (International Conference on Harmonisation of Technical Requirements for Registration of Pharmaceuticals for Human Use) guideline for testing linearity in validation of analytical procedures suggests that “linearity should be evaluated by visual inspection of a plot of signals as a function of analyte concentration or content.” The EP6‐A guideline recommends more quantitative methods that compare straight‐line and higher‐order polynomial curve fits. In this paper, a new equivalence test is proposed to compare the quality of a straight‐line fit to that of a higher‐order polynomial. By using orthogonal polynomials and generalized pivotal quantity analysis, one may estimate the probability of equivalence between a straight line and a polynomial curve fit either in the assay signal space (the Y values) or in the concentration space (the X values). In the special case of the linear‐to‐quadratic polynomial comparison, an equivalence test may be constructed via a two one‐sided T test. Copyright © 2013 John Wiley & Sons, Ltd.     

PHYSICAL PROPERTIES OF SOME NOVEL NEMATIC BIMESOGENS FOR THE FLEXOELECTRIC EFFECT

Abstract

In a chiral nematic liquid crystal, the flexoelectric effect consists of a fast and linear coupling with an applied electric field. One difficulty to overcome is the unwinding of the helix that occurs at higher fields due to dielectric coupling. The use of bimesogens, which possess very low molecular dielectric anisotropy can improve flexoelectric characteristics. New bimesogen compounds have recently been synthesised that exhibit switching angles of 45° for applied fields of about 9 V.μm^?1. In this paper, results from dielectric, electro-optic and dynamic light scattering measurements are reported for the new bimesogenic mixture. The dielectric anisotropy Δε changes sign with temperature and its values range between -0.2 and 0.3 for the temperature range studied. For Δε weakly positive, no electric field Freedericksz transition could be induced but Williams domains are observed instead. The large decrease in the bend elastic constant to viscosity coefficient ratio is attributed to a large increase in the bend viscosity coefficient.     

A dynamical systems approach to mixing and segregation of granular materials in tumblers

The physics of granular matter is one of the big questions in science. Granular matter serves as a prototype of collective systems far from equilibrium and fundamental questions remain. At the same time, an understanding of granular matter has tremendous practical importance. Among practical problems, granular mixing and its interplay with segregation is arguably at the top of the list in terms of impact. Granular mixing in three-dimensional systems is complicated, as flow induces segregation by particle size or density. Several approaches and points of view for analysis are possible in principle, ranging from continuum to discrete. Flow and segregation in three-dimensional systems is seemingly complicated; however, to a reasonable approximation, all of the dynamics takes place in a thin flowing surface layer. This observation, coupled with key experimental results, leads to a simple, compact and extensible continuum-based dynamical systems framework applicable to time-periodic flow in quasi-two-dimensional tumblers and three-dimensional systems (such as spheres and cubes) rotated about one or more axes of rotation. The case of time-periodic systems, in its simplest version, can be viewed as a mapping of a domain into itself. The placement of periodic points can be investigated using symmetry concepts; the character of the periodic points and associated manifolds provides a skeleton for the flow and a template for segregation processes occurring in the flow.     

Meeting Reports: Energy-Recovering Linacs for Next-Generation Light Sources

Energy-recovering linear accelerators (ERLs) have the potential of providing two orders of magnitude increase in brightness over third-generation sources while offering bunch lengths as low as tens of femtoseconds for timing experiments. ERLs are different from storage rings in that the electron bunches circulate once only. Thus they are not at equilibrium, and in particular can be packed much more tightly than in storage rings, resulting in brighter sources. However, dumping a 100 mA, 3 GeV beam that carries 300 MW of energy is not practical, so the energy is recovered, and the devices are in some sense energystorage rings.     

Sample size for the Z test and its confidence interval

The statistical power of a significance test is closely related to the length of the confidence interval (i.e. estimate precision). In the case of a Z test, the length of the confidence interval can be expressed as a function of the statistical power.     

An Approximation Method for Improving Dynamic Network Model Fitting

There has been a great deal of interest recently in the modeling and simulation of dynamic networks, that is, networks that change over time. One promising model is the separable temporal exponential-family random graph model (ERGM) of Krivitsky and Handcock, which treats the formation and dissolution of ties in parallel at each time step as independent ERGMs. However, the computational cost of fitting these models can be substantial, particularly for large, sparse networks. Fitting cross-sectional models for observations of a network at a single point in time, while still a nonnegligible computational burden, is much easier. This article examines model fitting when the available data consist of independent measures of cross-sectional network structure and the duration of relationships under the assumption of stationarity. We introduce a simple approximation to the dynamic parameters for sparse networks with relationships of moderate or long duration and show that the approximation method works best in precisely those cases where parameter estimation is most likely to fail—networks with very little change at each time step. We consider a variety of cases: Bernoulli formation and dissolution of ties, independent-tie formation and Bernoulli dissolution, independent-tie formation and dissolution, and dependent-tie formation models.