Rehabilitation of the Gauss-Jordan algorithm

Summary In this paper a Gauss-Jordan algorithm with column interchanges is presented and analysed. We show that, in contrast with Gaussian elimination, the Gauss-Jordan algorithm has essentially differing properties when using column interchanges instead of row interchanges for improving the numerical stability. For solutions obtained by Gauss-Jordan with column interchanges, a more satisfactory bound for the residual norm can be given. The analysis gives theoretical evidence that the algorithm yields numerical solutions as good as those obtained by Gaussian elimination and that, in most practical situations, the residuals are equally small. This is confirmed by numerical experiments. Moreover, timing experiments on a Cyber 205 vector computer show that the algorithm presented has good vectorisation properties.     

Fenchel duality, Fitzpatrick functions and the extension of firmly nonexpansive mappings

Recently, S. Reich and S. Simons provided a novel proof of the Kirszbraun-Valentine extension theorem using Fenchel duality and Fitzpatrick functions. In the same spirit, we provide a new proof of an extension result for firmly nonexpansive mappings with an optimally localized range.

     

Particle Characterization by Projected Area Determination

The projected areas of non-spherical particles do not represent an unambiguous particle characteristic. Depending on the orientation towards a constant observational direction, different projected areas result. The spectrum of all projected area values of a particle, if determined representatively, gives the probability with which a certain value is obtained by a single measurement. In this work, the frequency distributions of different examples of test objects were both calculated and measured. The objects were a cube, a rectangular parallelepiped and also three model agglomerates consisting of spheres of the same size. Instead of just one projected area, during each measuring procedure three projected areas from three orthogonal directions can be obtained. A mean value is then calculated to reduce the ambiguity of the particle characteristic and enhance the resolution. A suitable measurement set-up is introduced. The results of calculation and measurement are compared for observation from just one direction and also simultaneous observation from three directions. The frequency distributions of the equivalent diameters of the particle projected areas show a characteristic trend of the total curve with remarkable properties. The simultaneous measurement of three values from mutually orthogonal directions and their mean value calculation result in a much narrower distribution. In this case, a non-sphericity factor can additionally be calculated, whose frequency distribution contains information in a characteristic manner about the degree to which the particle shape differs from a sphere.