A multiplicatively-weighted Voronoi diagram approach to logistics districting

The objective of districting in logistics distribution problems is to find a near optimal partition of the served region into delivery zones or districts. This kind of problem has been treated in the literature with geometric-shaped districts and continuous approximations. Usually it is assumed an underlying road network equivalent to a Euclidean, rectangular, or ring-radial metric. In most real problems, however, the road network is a coarse combination of metrics. In this context, it is unclear what is the optimal shape of the districts and how one should orientate them. A possibility is to treat the problem with a Voronoi diagram approach. Departing from a previously determined ring-radial districting pattern and relaxing the initial district boundaries, we apply the multiplicatively-weighted Voronoi diagram formulation in order to smooth district contours. The computing process is iterated until the convergence is attained. The model has been applied to solve a parcel delivery problem in the city of São Paulo, Brazil, whose results are presented and discussed in the paper.     

Parallelizing Feature Selection

Classification is a key problem in machine learning/data mining. Algorithms for classification have the ability to predict the class of a new instance after having been trained on data representing past experience in classifying instances. However, the presence of a large number of features in training data can hurt the classification capacity of a machine learning algorithm. The Feature Selection problem involves discovering a subset of features such that a classifier built only with this subset would attain predictive accuracy no worse than a classifier built from the entire set of features. Several algorithms have been proposed to solve this problem. In this paper we discuss how parallelism can be used to improve the performance of feature selection algorithms. In particular, we present, discuss and evaluate a coarse-grained parallel version of the feature selection algorithm FortalFS. This algorithm performs well compared with other solutions and it has certain characteristics that makes it a good candidate for parallelization. Our parallel design is based on the master--slave design pattern. Promising results show that this approach is able to achieve near optimum speedups in the context of Amdahl's Law.     

Continuum modelling and numerical simulation of material damage at finite strains

Summary This paper describes in detail a general framework for the continuum modelling and numerical simulation of internal damage in finitely deformed solids. The development of constitutive models for material deterioration is addressed within the context of Continuum Damage Mechanics. Links between micromechanical aspects of damage and phenomenological modelling within continuum thermodynamics are discussed and a brief historical review of Continuum Damage Mechanics is presented. On the computational side, an up-to-date approach to the finite element solution of large strain problems involving dissipative materials is adopted. It relies on an implicit finite element discretization set on the spatial configuration in conjunction with the full Newton-Raphson scheme for the iterative solution of the corresponding non-linear systems of equations. Issues related to the numerical integration of the path dependent damage constitutive equations are discussed in detail and particular emphasis is placed on the consistent linearization of associated algorithms. A model for elastic damage in polymers and finite strain extensions to Lemaitre's and Gurson's models for ductile damage in metals are formulated within the described framework. The adequacy of the constitutive-numerical framework for the simulation of damage in large scale industrial problems is demonstrated by means of numerical examples.     

Static output feedback controllers: stability and convexity

The main objective of this paper is to solve the following stabilizing output feedback control problem: given matrices (A; B₂ ; C₂) with appropriate dimensions, find (if one exists) a static output feedback gain L such that the closed-loop matrix A-B₂LC₂ is asymptotically stable. It is known that the existence of L is equivalent to the existence of a positive definite matrix belonging to a convex set such that its inverse belongs to another convex set. Conditions are provided for the convergence of an algorithm which decomposes the determination of the aforementioned matrix in a sequence of convex programs. Hence, this paper provides a new sufficient (but not necessary) condition for the solvability of the above stabilizing output feedback control problem. As a natural extension, we also discuss a simple procedure for the determination of a stabilizing output feedback gain assuring good suboptimal performance with respect to a given quadratic index. Some examples borrowed from the literature are solved to illustrate the theoretical results     

The evaluation of the use of DEAE-dextran in glycemic control in diabetic patients in pregnancy

Thirty pregnant women with a pre-gestational history of type II diabetes or sugar intolerance and recruited during the second trimester of pregnancy, were administered DEAE-dextran (1 g x 3 times a day) in association with compensatory insulin therapy. Results of the end of trial tests showed amelioration of all of the parameters studied. The sugar curve after 120' from glucose load (100 g at fasting), showed a highly significant decrease at T90. Triglycerides T0 vs T90 gave p = 0.0001, probably due to improved body utilization of the insulin. DEAE-dextran was well tolerated and all of the patients enrolled at the beginning, completed the trial.     

Thermal restraint and fire resistance of columns

A proposal is made, based on the results of a series of tests and calculations, with the aim of being used as a simple method to correct the value of the critical temperature of steel columns free to elongate, in order to take into account the restraint effect of the structure to which they belong in a practical situation. To better illustrate the possible types of behaviour of heated steel columns with elastic restraint to the thermal elongation, and the reasons why the critical temperature of axially loaded slender steel columns with thermal restraint can sometimes be lower than the critical temperature of the same columns free to elongate, a simple model is presented and used in a qualitative analysis.     

Free vibrations of orthotropic spherical shells

The equations of motion for the free vibrations of orthotropically stiffened open spherical shells are developed and solved using a suitable finite difference model. Variations in both flexural and extensional orthotropic stiffness properties are investigated by means of carefully selected parametric studies. A systematic examination of the contribution to strain energy in each mode, arising from the various components of orthotropic shell stiffness, is shown to assist the interpretation of the effects of orthotropic stiffness changes, and to allow prediction of approximate frequency spectra. Based on the analysis of a related isotropic spherical shell it is shown how a modified form of Rayleigh's method provides approximations of frequency spectra sufficiently accurate to assist the conceptual dynamic design process.     

Numerical analysis of nonlinear bistable optical waveguides

Numerical solutions illustrating the onset of bistability and hysteresis are presented for the symmetrical step index, asymmetrical step index, and asymmetrical diffused slab nonlinear optical waveguides. Two different numerical techniques have been used independently-a finite-element method and a variational method. Both methods produce numerically stable solutions, and agreement between them is good for both increasing and decreasing total power from below or above the threshold power. The results are compared with analytical solutions for these structures. The onset and the end of the physically unstable solutions regime coincides with the two power thresholds for the increasing and decreasing powers