OPUS 4 | Search

Demand forecasting for companies with many branches, low sales numbers per product, and non-recurring orderings (2006)

Sascha Kurz Jörg Rambau

We propose the new Top-Dog-Index to quantify the historic deviation of the supply data of many small branches for a commodity group from sales data. On the one hand, the common parametric assumptions on the customer demand distribution in the literature could not at all be supported in our real-world data set. On the other hand, a reasonably-looking non-parametric approach to estimate the demand distribution for the different branches directly from the sales distribution could only provide us with statistically weak and unreliable estimates for the future demand.

Proof-of-Concept Application - Annual Report Year 2 (2006)

Oscar Ardaiz Pablo Chacin Isaac Chao Juan Carlos Cruellas Felix Freitag Manuel Medina Leandro Navarro Miguel Valero Loviu Joita Omer F. Rana Torsten Eymann

Performance Evaluation - Annual Report Year 2 (2006)

Oscar Ardaiz Michele Catalano Pablo Chacin Isaac Chao Juan Carlos Cruellas Felix Freitag Manuel Medina Leandro Navarro Miguel Valero Liviu Joita Omer F. Rana Björn Schnizler Torsten Eymann

In this paper a performance measuring infrastructure,developed for the prototype and simulator, concering the experiment configuration, data measurement, and data collection, is presented. A corresponding performance evaluation framework is defined to obtain the metrics from the measured data. Initial experiments were carried out to test the developed prototype, simulator and the performance measuring infrastructure.

Simulator Development - Annual Report Year 2 (2006)

Gaetano Calabrese Björn Schnizler Werner Streitberger Torsten Eymann Floriano Zini

In this paper the simulation environment for the CATNETS project is defined further. The chosen simulator is adopted in terms of new features an architecture changes in order to provide a valid simulation environment for Application Layer Network scenarios. Furthermore the requirements for a scenario generator and the needed configuration mechanisms for the actual simulation runs are introduced.

Theoretical and Computational Basis for CATNETS - Annual Report Year 2 (2006)

Torsten Eymann Werner Streitberger Daniel Veit Georg Buss Björn Schnizler Dirk Neumann

In this work the self-organising potential of the CATNETS allocation mechanism is described to provide a more comprehensive view on the research done in this project. The formal description of either the centralised and decentralised approach is presented. Furthermore the agents’ bidding model is described and a comprehensive overview on how the catallactic mechanism is incorporated into the middleware and simulator environments is given.

A Theoretical and Computational Basis for CATNETS (2006)

Björn Schnizler Dirk Neumann Daniel Veit Michael Reinicke Werner Streitberger Torsten Eymann Felix Freitag Isaac Chao Pablo Chacin

The main content of this report is the identification and definition of market mechanisms for Application Layer Networks (ALNs). On basis of the structured Market Engineering process, the work comprises the identification of requirements which adequate market mechanisms for ALNs have to fulfill. Subsequently, two mechanisms for each, the centralized and the decentralized case are described in this document. These build the theoretical foundation for the work within the following two years of the CATNETS project.

The Use of Biomarker and Stable Isotope Analyses in Palaeopedology / Reconstruction of Middle and Late Quaternary Environmental and Climate History, with examples from Mt. Kilimanjaro, NE Siberia and NE Argentina (2006)

Michael Zech

Palaeosols are important terrestrial archives for the reconstruction of the Quaternary landscape and climate history. In order to derive reliable information from these archives about sedimentation, vegetation and climate history, various methods and proxies are traditionally applied, e.g. texture analysis, numeric dating methods and mineral analysis. The aim of this dissertation is to evaluate the potential of biomarker and stable isotope analyzes. Specifically, I focused on plant leaf wax-derived n-alkanes, amino acid enantiomers, stable carbon and nitrogen isotopes (d13C and d15N) in bulk soil organic matter (SOM) and on compound-specific isotope analysis (CSIA) of n-alkanes. The respective methods were partly optimized and then applied in multi-proxy analytical approaches to three selected palaeosol records, representing different ecological environments. In all three study areas, long-chain n-alkane ratios, nC31/nC27 and (nC31+nC29)/nC27, respectively, proved to be straightforward biomarker proxies for the reconstruction of the terrestrial palaeovegetation at plant community level (especially grasses and herbs versus trees). Short- and mid-chain n-alkanes (nC17–nC19 and nC20–nC25, respectively) were successfully used for detecting algal- and aquatic macrophyte-derived organic matter (OM) in the sediment core Arg. D4. Amino acid enantiomers as nitrogen (N) biomarkers allowed a further characterization of the SOM in the Tumara Palaeosol Sequence: On the one hand, the depth functions of D/L-aspartic acid (Asp) and D/L-lysine (Lys) could be roughly described by exponential fits, reflecting SOM aging. On the other hand, brown interglacial/-stadial palaeosols generally revealed higher D/L-ratios than dark gray glacial palaeosols. This finding suggests that D/L-aspartic acid and D/L-lysine may serve as palaeotemperature proxies. In the Arg. D4 record, d13C varied in a wide range (from -30.1‰ to -17.4‰), indicating C3-C4 vegetation changes and hence allowing a reconstruction of the palaeovegetation. However, the natural abundance of 13C was no straightforward proxy for the interpretation of the palaeosol sequences on Mt. Kilimanjaro and in the Tumara Valley. There, the interpretation of smaller d13C variations – assumed to be independent of C3-C4 vegetation changes – needed multi-proxy analytical approaches for disentangling the various possibly influencing environmental factors: On Mt. Kilimanjaro, d13C is higher in palaeosols, which developed under ericaceous vegetation (~ -25‰) compared to those developed under tropical montane forests (~ -27‰), suggesting that such vegetation changes are responsible for the observed d13C pattern. In the Tumara Palaeosol Sequence, d13C correlates negatively with total organic carbon (TOC) and TOC/N. As both parameters may serve as proxies for SOM decomposition, it is assumed that degradation processes have contributed significantly to this d13C record. Furthermore, also changing water stress conditions for the plants could have played a crucial role for d13C in the Tumara Palaeosol Sequence. In contrast to d13C, d15N in the Tumara Palaeosol Sequence does not correlate with any of the other SOM characterizing parameters (TOC, TOC/N and d13C). Although other processes than SOM decomposition like (i) denitrification, (ii) N fixation, (iii) N losses by frequent fire events, and (iv) changes in the atmospheric 15N deposition are discussed as factors contributing to an open N cycle, d15N in the Tumara Palaeosol Sequence seems not to be a straightforward proxy. The compound-specific d13C analysis (CSIA) of n-alkanes was optimized and applied to selected samples from the Arg. D4 record. The highly significant correlations of the compound-specific isotope results with bulk d13C corroborate the reliability of the d13C vegetation proxy. Furthermore, the increasing d13C amplitudes from nC27 to nC33 validate the origin of these biomarker molecules, with nC27 and nC29 mainly deriving from C3 trees and shrubs and nC31 and nC33 mainly deriving from C3 or C4 grasses and herbs. Eventually, the multi-proxy analytical approaches (including the innovative biomarker and stable isotope analyzes) enabled detailed reconstructions of the Middle and Late Quaternary palaeoenvironmental changes in the three study areas: Accordingly, the deep black palaeosols on the southern slopes of Mt. Kilimanjaro reflect periods of climatic deterioration during the Last Glacial Maximum (LGM) and the Late Glacial, which coincided with a descent of the ericaceous vegetation belt. The palaeopedologic findings from the Tumara Palaeosol Sequence suggest that the dark gray and brown stratigraphic units of this record describe alternating glacial and interglacial/-stadial periods during the last ~240 ka. The stratigraphic units of the Arg. D4 record were correlated with climatic events on the Bolivian Altiplano and discussed in terms of an intensified palaeo-South American Summer Monsoon (SASM).

The Corticolous Crustose and Microfoliose Lichens of Northeastern Brazil – Diversity, Ecology, and Conservation (2006)

Marcela Eugenia Da Silva Cáceres

Collections of crustose and microfoliose corticolous lichens made in a number of 22 localities of Atlantic rainforest, Caatinga and Brejos de Altitudes (caatinga enclaves) in the states of Rio Grande do Norte, Paraiba, Pernambuco, Alagoas, and Sergipe, in northeastern Brazil yielded a total of 450 species, distributed in 110 genera, in 32 families, 12 orders, and 4 classes of Ascomycota and one of Basidiomycota. A total of 14 taxa are described here as new to science in the genera Aciculopsora, Bacidina, Calopadia, Cryptothecia, Enterographa, Graphis, Malcolmiella, Phaeographis, and Plectocarpon. In addition, 5 new combinations are proposed in the genera Chapsa and 7 in the genera Malcolmiella. Identification keys to the genera and species of corticolous crustose and microfoliose lichens of northeastern Brazil are provided, with complete checklist and descriptions of new species. The comparison between different vegetation types and localities across the study area used the lichen species composition at each site to ordinate and classify sites. The highest dissimilarity was registered between the Atlantic rainforest sites, with an average of 21% and maximum of 55%. The lichen species composition from the Atlantic rainforest sites as a whole compared to the Caatinga sites showed dissimilarity values averaging 0.92 or 8% of shared species. The influence of tree bark characteristics and phorophyte species on lichen species composition, richness, and area cover, in a selected fragment of Atlantic rainforest was analyzed. Multivariate analysis of sample plots, including non-metric multidimensional scaling (NMS), detrended correspondence analysis (DCA), and canonical correspondence analysis (CCA), and also cluster analysis, indicated subtle patterns of phorophyte preferences among certain lichen species, as well as correlation with environmental parameters, in particular bark pH, degree of bark shedding, and density and size of bark lenticels. Individual and multiple correlation also revealed correlations between lichen species richness and area cover on one hand and bark pH (negative), density and size of bark lenticels (negative), degree of bark shedding (negative), presence of milk sap (positive), and diffuse site factor (positive). No distinct lichen communities were detected among the samples, but cluster analysis revealed three main sample groups and six subgroups with slightly different lichen species composition, each one with characteristic indicator species but with highly variable overall species composition. It is concluded that community formation in tropical rainforest understory lichens is governed by two main factors, phorophyte bark characteristics and microclimate, but is largely obscured by the stochastic effects of species dispersal and rare species, and also the amount of phorophyte tree diversity. It is predicted that phorophyte specificity is best observed in model systems with low tree and low lichen diversity, distinct differences between tree species in terms of bark characteristics, homogeneous population structure, and low microclimatic variation. Finally, three different sampling methods were considered on the present study and their efficiency for more accurate estimation of tropical microlichen diversity was tested. The analysis showed that opportunistic sampling fails to detect rare, inconspicuous, sterile, and/or cryptic species, usually neglected or overlooked in the field. It is also apparent that it is not the higher number of specimens collected via quantitative sampling that results in a higher number of species, but the method of selection of the specimens, which is subjective and biased towards abundant, conspicuous, fertile and/or distinctive species in opportunistic sampling, but objective and unbiased in quantitative sampling.

Source identification of trichloroacetic acid with preparative capillary gas chromatography and accelerator mass spectrometry (2006)

Kamila Jakubowska-Switaj

A method for radiocarbon source identification of environmental chemicals present in soil at ppb levels using trichloroacetic acid (TCA) as an exemplary compound was developed in the present thesis. A powerful approach to differentiate between biogenic and anthropogenic sources of environmental chemicals is to determine the radiocarbon content: an organic chemical derived from the biological carbon pool, and thus resulting from assimilation of atmospheric carbon dioxide, contains the radioactive carbon isotope 14C at the typical atmospheric equilibrium abundance of 14C/12C = 1.2 x 10-12, while a chemical derived from fossil carbon over 200 million years old is free of 14C (half life 5730 years). The only method being sensitive enough for direct quantitative determination of a few thousand atoms of radiocarbon in a minute amount of a pure environmental trace chemical is accelerator mass spectrometry (AMS). In the present thesis an AMS equipped with a titanium-target gas source was used. Therefore, 1 microgram carbon in the form of carbon dioxide was sufficient for the measurement. To determine TCA in soil the total amount of soil sample needed depends on its concentration in this compartment. For sampling, sites of the Bavarian spruce site observational network were selected where TCA concentrations were relatively high. Two sampling sites were compared and the one with the higher TCA soil burden was chosen. Soil samples from Oberthulba were collected, and TCA was isolated in sufficient quantity in pure form, by a combination of differential liquid extraction steps, followed by derivatization, and preparative capillary-gas chromatography (PC-GC). In environmental samples the actual analyte constitutes usually only a minute fraction, mostly less than 0.1 % of the primary extract. Taking this into consideration, sequential (two-dimensional) PC-GC-separation on two different columns was necessary emphasizing high chromatographic capacity and high stationary phase selectivity in the first step and high separation efficiency in the second one. This was achieved by using two columns of appropriate dimensions. In the first dimension a wide-bore high-capacity thick film capillary coated with a non-polar stationary phase was used. For further removal of by-products the fraction containing the analyte was cryogenically collected, and the final purity refinement was effected on a short thin-film, high-efficiency capillary coated with a medium-polar stationary phase. From a sequence of runs, the required amount of trichloroacetic acid methyl ester (TCA-OMe) was collected with a cryogenic fraction collector, combusted to carbon dioxide, water and chlorine were separated, and carbon dioxide was fed into the gas ionization source of the AMS-instrument. Great care was taken to avoid blank contamination during sample preparation steps for AMS because at such low levels significant complicating factors are the ubiquitous chemical trace pollutants in laboratory air and the presence of atmospheric carbon dioxide containing natural occurring 14C. The radiocarbon determination of a real sample of TCA collected in the mineral horizon of forest soil in Northern Bavaria yielded a value of 2.2 percent modern carbon (PMC) for TCA-OMe corresponding to 3.3 PMC for TCA indicating the anthropogenic origin of TCA. If TCA were of biogenic origin, a value of 66.7 PMC would be expected as one additional carbon from a petrochemical source is introduced during derivatization to the methyl ester. The result proved that the TCA found in the soil at the particular sampling location was an atmospheric breakdown product of tetrachloroethylene and not a product of any purported soil microbial process. This result is important for risk assessment procedures performed for TCA and tetrachloroethylene under OECD and EU chemical programs. To control emissions of environmental chemicals and to reduce their influence on the various environmental compartments it must be known if they originate from anthropogenic sources.

Vibrated granular matter (2006)

Andreas Götzendorfer

Granular matter is defined as a large collection of particles the size of which is larger than one micron so that Brownian motion is negligible. Its behavior has been studied at least since the days of Charles-Augustin de Coulomb (1736-1806), who originally stated his law of friction for granular materials. In the physics community interest in granular media started to grow considerably around 1990, driven by the fast-growing performances of computer simulations. Since then the number of publications in this field has surged enormously. Because of the dissipative nature of particle collisions, in order to maintain a steady flow or a dynamic steady state, energy has to be fed constantly into a granular system. In lab experiments this is often done by applying a sinusoidal horizontal or vertical oscillation to the container. One of the aims of this work was to study effects of the combined action of both forms of agitation. In the presented experiments vertical and horizontal oscillations were superposed such that every point of the support followed a circular trajectory. By choosing a ring-shaped container geometry, the long-time dynamics of a closed, mass conserving system devoid of disturbances from the influx and outpouring of grains could be studied. This setup was used to examine spatially extended surface wave patterns of a granular bed. Standing waves oscillating at half the forcing frequency were observed within a certain range of the driving acceleration. The dominant wavelength of the pattern was measured for various forcing frequencies at constant amplitude. These waves are not stationary, but drift with a velocity equal to the transport velocity of the granular material, determined by means of a tracer particle. At higher forcing strength localized period doubling waves arise. These traveling solitary wave packets are accompanied by a locally increased particle density. The length and velocity of the granular wave pulse were measured as a function of the amount of material in the container. Inspired by traffic flow models that explain the spontaneous appearance of pulses – “phantom jams” - out of initially homogeneous flow a simple continuum model for the material distribution was developed. Based on the measured granular transport velocity as a function of the bed thickness, it captures the essence of the experimental findings. Furthermore the fluidization of a monolayer of circularly vibrated glass beads was studied. At peak forcing accelerations within a certain interval a solid-like and a gas-like domain coexist. The solid fraction decreases with increasing acceleration and shows hysteresis. Complementary to the experimental studies a molecular dynamics simulation was used to extract local granular temperature, basically defined as the variance of the particle velocity distribution, and number density. It was found that the number density in the solid phase is several times that in the gas, while the temperature is orders of magnitude lower. To investigate the transition of a crystalline particle packing to a fully fluidized state a separate setup was used. Particles were confined to two dimensions in order to keep them visible at all times. With the help of a high speed camera all particles could then be traced. The vibration was restricted to the vertical direction. The experiment was designed flexible enough to allow an easy variation of driving parameters and the use of particles of various sizes. An initially close packed granular bed was exposed to sinusoidal container oscillations with gradually increasing amplitude. At first the particles close to the free surface become mobile. When a critical value of the forcing strength is reached the remaining crystal suddenly breaks up and the bed fluidizes completely. This transition leads to discontinuous changes in the density distribution and in the root mean square displacement of the individual particles. Likewise the vertical center of mass coordinate increases by leaps and bounds at the transition. It turns out that the maximum container velocity v is the crucial driving parameter determining the state of a fully fluidized system. For particles of various sizes the transition to full fluidization occurs at the same value of v^2/gd, where d is the particle diameter and g is the gravitational acceleration.

Refine

Author

Year of publication

Document Type

Language

Keywords

Institute

28 search hits