DTU Orbit RSS Feed
http://orbit.dtu.dk/en/publishers-research/technical-university-of-denmark-dtu(f7c040ab-194a-4da4-b458-670da24b2a08)/publications.rss?filter=research
RSS FeedAlgorithms and data structures for grammar-compressed strings
http://orbit.dtu.dk/en/publications/algorithms-and-data-structures-for-grammarcompressed-strings(78d89068-7c76-4105-8908-a2d474683732).html
<div style='font-size: 9px;'><div class="rendering rendering_publication rendering_publication_short rendering_bookanthology rendering_short rendering_bookanthology_short"><h2 class="title"><a class="link" rel="BookAnthology" href="http://orbit.dtu.dk/en/publications/algorithms-and-data-structures-for-grammarcompressed-strings(78d89068-7c76-4105-8908-a2d474683732).html"><span>Algorithms and data structures for grammar-compressed strings</span></a></h2><a class="link person" rel="Person" href="http://orbit.dtu.dk/en/persons/patrick-hagge-cording(5bccf823-af4b-47b6-b1c6-2d45a362a261).html"><span>Cording, P. H.</span></a>, <a class="link person" rel="Person" href="http://orbit.dtu.dk/en/persons/philip-bille(81ad6edf-aa11-47bc-adaf-227fe7e6963e).html"><span>Bille, P.</span></a> & <a class="link person" rel="Person" href="http://orbit.dtu.dk/en/persons/inge-li-goertz(08da98e0-b1f4-492e-94e7-15083cc40701).html"><span>Gørtz, I. L.</span></a> <span class="date">2014</span> Kgs. Lyngby: <a class="link" rel="Publisher" href="http://orbit.dtu.dk/en/publishers/technical-university-of-denmark-dtu(f7c040ab-194a-4da4-b458-670da24b2a08).html"><span>Technical University of Denmark (DTU)</span></a>. <span class="numberofpages">66 p.</span> (DTU Compute PHD-2014; No. 357).<p class="type"><span class="type_family">Publication<span class="type_family_sep">: </span></span><span class="type_classification_parent">Research<span class="type_parent_sep"> › </span></span><span class="type_classification">Ph.D. thesis – Annual report year: 2014</span></p></div><div class="rendering rendering_publication rendering_publication_detailsportal rendering_bookanthology rendering_detailsportal rendering_bookanthology_detailsportal"><div class="phddissertation"><table class="properties"><tbody><tr class="language"><th>Original language</th><td>English</td></tr><tr><th>Publication date</th><td><span class="date">2014</span></td></tr></tbody></table><table class="properties"><tbody><tr><th>Place of publication</th><td>Kgs. Lyngby</td></tr><tr><th>Publisher</th><td><a class="link" rel="Publisher" href="http://orbit.dtu.dk/en/publishers/technical-university-of-denmark-dtu(f7c040ab-194a-4da4-b458-670da24b2a08).html"><span>Technical University of Denmark (DTU)</span></a></td></tr><tr><th>Number of pages</th><td>66</td></tr><tr class="status"><th>State</th><td>Submitted</td></tr></tbody></table></div> <table class="properties"><tbody><tr><th>Name</th><td>DTU Compute PHD-2014</td></tr><tr><th>Number</th><td>357</td></tr><tr><th>ISSN (print)</th><td>0909-3192</td></tr></tbody></table></div></div>Tue, 31 Dec 2013 23:00:00 GMThttp://orbit.dtu.dk/en/publications/algorithms-and-data-structures-for-grammarcompressed-strings(78d89068-7c76-4105-8908-a2d474683732).html2013-12-31T23:00:00ZAlgorithms for Electromagnetic Scattering Analysis of Electrically Large Structures
http://orbit.dtu.dk/en/publications/algorithms-for-electromagnetic-scattering-analysis-of-electrically-large-structures(68a87a2a-3b48-4d25-aac3-613f22505e72).html
<div style='font-size: 9px;'><div class="rendering rendering_publication rendering_publication_short rendering_bookanthology rendering_short rendering_bookanthology_short"><h2 class="title"><a class="link" rel="BookAnthology" href="http://orbit.dtu.dk/en/publications/algorithms-for-electromagnetic-scattering-analysis-of-electrically-large-structures(68a87a2a-3b48-4d25-aac3-613f22505e72).html"><span>Algorithms for Electromagnetic Scattering Analysis of Electrically Large Structures</span></a></h2><a class="link person" rel="Person" href="http://orbit.dtu.dk/en/persons/oscar-peter-borries(07ed5fe0-b447-49a9-9fb9-5555c76d072d).html"><span>Borries, O. P.</span></a> & <a class="link person" rel="Person" href="http://orbit.dtu.dk/en/persons/per-christian-hansen(6c6c6f32-e3b2-40d5-a407-aa8e2f051f7b).html"><span>Hansen, P. C.</span></a> <span class="date">2014</span> Kgs. Lyngby: <a class="link" rel="Publisher" href="http://orbit.dtu.dk/en/publishers/technical-university-of-denmark-dtu(f7c040ab-194a-4da4-b458-670da24b2a08).html"><span>Technical University of Denmark (DTU)</span></a>. <span class="numberofpages">205 p.</span> (DTU Compute PHD-2014; No. 354).<p class="type"><span class="type_family">Publication<span class="type_family_sep">: </span></span><span class="type_classification_parent">Research<span class="type_parent_sep"> › </span></span><span class="type_classification">Ph.D. thesis – Annual report year: 2014</span></p></div><div class="rendering rendering_publication rendering_publication_detailsportal rendering_bookanthology rendering_detailsportal rendering_bookanthology_detailsportal"><div class="textblock">Accurate analysis of electrically large antennas is often done using either Physical Optics (PO) or Method of Moments (MoM), where the former typically requires fewer computational resources but has a limited application regime. This study has focused on fast variants of these two methods, with the goal of reducing the computational complexity while maintaining accuracy. <br /><br />Regarding MoM, the complexity is reduced by applying the Multi-Level Fast Multipole Method (MLFMM) in combination with an iterative solver. Using MLFMM with a MoM implementation based on Higher-Order (HO) basis functions has, by several authors, been dismissed as being too memory intensive. In the present work, we demonstrate for the first time that by including a range of both novel and previously presented modifications to the standard MLFMM implementation, HO MLFMM can achieve both memory reduction and significant speed increase compared to Lower-Order (e.g., RWG) based MLFMM. Further, issues surrounding an iterative solution, such as the iterative solver and preconditioning, are discussed. Numerical results demonstrate the performance and stability of the algorithm for very large problems, including full satellites at Ku band.<br /><br />Accelerating PO is an entirely different matter. A few authors have discussed applying the Fast-PO technique to far fields, achieving relative errors of 0.1%−1% for moderately sized scatterers. For near-fields, the state-of-the-art implementation of Fast-PO has several difficulties, in particular low accuracy and limited application regime. For the problems considered in this thesis, the error limit for PO is ≈ 0.01%, and the application limitations of the published Fast-PO are too prohibitive for our use. Therefore, results based on an improved Fast-PO implementation for far-fields, as well as a novel algorithm for near-fields, are presented. These results demonstrate that it is possible to achieve very accurate results, with relative errors around 10<sup>−5</sup>, at a much reduced time consumption. The method behind this part of the code is deemed confidential by TICRA.<br /></div><div class="phddissertation"><table class="properties"><tbody><tr class="language"><th>Original language</th><td>English</td></tr><tr><th>Publication date</th><td><span class="date">2014</span></td></tr></tbody></table><table class="properties"><tbody><tr><th>Place of publication</th><td>Kgs. Lyngby</td></tr><tr><th>Publisher</th><td><a class="link" rel="Publisher" href="http://orbit.dtu.dk/en/publishers/technical-university-of-denmark-dtu(f7c040ab-194a-4da4-b458-670da24b2a08).html"><span>Technical University of Denmark (DTU)</span></a></td></tr><tr><th>Number of pages</th><td>205</td></tr><tr class="status"><th>State</th><td>Submitted</td></tr></tbody></table></div> <table class="properties"><tbody><tr><th>Name</th><td>DTU Compute PHD-2014</td></tr><tr><th>Number</th><td>354</td></tr><tr><th>ISSN (print)</th><td>0909-3192</td></tr></tbody></table></div></div>Tue, 31 Dec 2013 23:00:00 GMThttp://orbit.dtu.dk/en/publications/algorithms-for-electromagnetic-scattering-analysis-of-electrically-large-structures(68a87a2a-3b48-4d25-aac3-613f22505e72).html2013-12-31T23:00:00ZFull Scale Measurements of the Hydro-Elastic Response of Large Container Ships for Decision Support
http://orbit.dtu.dk/en/publications/full-scale-measurements-of-the-hydroelastic-response-of-large-container-ships-for-decision-support(80015fe7-b181-4e59-97fd-61f330063fe3).html
<div style='font-size: 9px;'><div class="rendering rendering_publication rendering_publication_short rendering_bookanthology rendering_short rendering_bookanthology_short"><h2 class="title"><a class="link" rel="BookAnthology" href="http://orbit.dtu.dk/en/publications/full-scale-measurements-of-the-hydroelastic-response-of-large-container-ships-for-decision-support(80015fe7-b181-4e59-97fd-61f330063fe3).html"><span>Full Scale Measurements of the Hydro-Elastic Response of Large Container Ships for Decision Support</span></a></h2><a class="link person" rel="Person" href="http://orbit.dtu.dk/en/persons/ingrid-marie-vincent-andersen(817b53b7-4749-4650-93b8-48fe175ac288).html"><span>Andersen, I. M. V.</span></a>, <a class="link person" rel="Person" href="http://orbit.dtu.dk/en/persons/joergen-juncher-jensen(7ba41725-e940-4ecd-9bf5-d25e44c9aabe).html"><span>Jensen, J. J.</span></a> & <a class="link person" rel="Person" href="http://orbit.dtu.dk/en/persons/ulrik-dam-nielsen(106115f2-5a14-4e9b-ba17-0199c93439b5).html"><span>Nielsen, U. D.</span></a> <span class="date">2014</span> Kgs. Lyngby: <a class="link" rel="Publisher" href="http://orbit.dtu.dk/en/publishers/technical-university-of-denmark-dtu(f7c040ab-194a-4da4-b458-670da24b2a08).html"><span>Technical University of Denmark (DTU)</span></a>. <span class="numberofpages">416 p.</span> (DCAMM Special Report; No. S166).<p class="type"><span class="type_family">Publication<span class="type_family_sep">: </span></span><span class="type_classification_parent">Research<span class="type_parent_sep"> › </span></span><span class="type_classification">Ph.D. thesis – Annual report year: 2014</span></p></div><div class="rendering rendering_publication rendering_publication_detailsportal rendering_bookanthology rendering_detailsportal rendering_bookanthology_detailsportal"><div class="textblock">The overall topic of this thesis is decision support for operation of ships and several aspects are covered herein. However, the main focus is on the wave-induced hydro-elastic response of large container ships and its implications on the structural response. <br /><br />The analyses are based mainly on full scale measurements from four container ships of 4,400 TEU, 8,600 TEU, 9,400 TEU and 14,000 TEU Primarily, strains measured near the deck amidships are used. Furthermore, measurements of motions and the encountered sea state are available for one of the ships. The smallest ship is in operation on the North Atlantic, while the three largest ships are operated on the Europe - Asia route. <br /><br />In the design rules of the classification societies for container ships the minimum design sagging bending moment amidships is larger than the hogging bending moment. Due to their design (full midship section and slender bow and stern sections) and their normal cargo loading condition, container ships are typically operated in a still-water hogging condition (tension in deck and compression in the bottom structure). The wave-induced bending moment is added to the still-water bending moment, which, together with the smaller design hogging bending moment, generally makes the wave-induced hogging bending moment more critical than the sagging bending moment in the operation of container ships.<br /><br />As container ships of today become larger, their natural vibration frequencies become lower and approach the typical encounter frequency with the waves. Together with the relatively high design speed and often pronounced bow flare this makes large container ship more sensitive to slamming and, consequently, the effects of wave-induced hull girder vibrations. From full scale strain measurements of individual, measured hull girder responses in the four container ships, the wave-induced hull girder vibrations are found to increase the vertical bending moment amidships by 100% or more. From the full scale measurements the amplification, due to the hull girder flexibility, is found to be largest for the 8,600 TEU and the 9,400 TEU ships, but, in addition to ship size, speed and<br />bow flare angle are also believed to be important factors contributing to the hull girder vibrations.<br /><br />The hull girder vibrational response is found to be dominated by the 2-node vertical bending mode. No torsional vibrations are found but torsion may, however, still be a concern for ultra large container ships. The damping of the 2-node vertical bending mode is estimated from full scale measurements for the four ships to 1.3-2.5% of the critical damping. No effect of ship size on the damping is identified.<br /><br />In some cases the hogging bending moment is more amplified by the effect of the hull girder flexibility than the sagging bending moment. No general trend in the amplification of the response is found from the full scale measurements. In some cases, the rigid-body hogging bending moment, found from full scale measurements and model tests, is considerably larger than the corresponding sagging bending moment. Generally, the difference between the design sagging and hogging bending moments is not reflected in the full scale measurements considered here.<br /><br />The extreme value of the vertical hogging bending moment, as estimated from full scale measurements, is investigated using the peak-over-threshold method for different periods.<br />The tails of the peak distributions for the four different ships are found to be very different from case to case. The irregularity of the tail behaviour makes it difficult to determine<br />an appropriate extreme value distribution for the hogging bending moment. The Gumbel distribution is believed to be the appropriate extreme value distribution, but it may be necessary to fit other types of extreme value distributions to the largest peaks.<br /><br />From the full scale measurements it is difficult to assess the influence of operational parameters (ship speed, heading relative to the waves and wave height) on the extreme response because these data are not readily available in all cases. Model tests indicate that bow-quartering sea may induce larger structural loads on the ship than direct head sea and that the amplification of the response due to the hull girder flexibility is largest in bow-quartering waves. However, this fact is not necessarily reflected in the behaviour of ship masters who seemingly tend to prefer bow-quartering sea to head sea when encountering adverse weather.<br /><br />Numerical design tools are widely used in ship design, but may not be able to fully capture the effect of the hull girder flexibility and are here found to significantly underestimate the effect compared to model tests and full scale measurements. Hence, full scale measurements from ships are highly valuable in the evaluation of existing designs and may reveal effects that cannot be assessed numerically.<br /><br />For decision support, accurate knowledge of the encountered sea state parameters (wave height, period and relative wave direction) is crucial. One means to estimate the on-site sea state from an advancing ship is to use the wave buoy analogy, i.e. use the motions of the ship and the associated motion transfer functions to derive the sea state parameters.<br />The method is promising but needs further refinement before it can be implemented in decision support systems on board ships.<br /><br />Fatigue damage is estimated from full scale strain measurements from two of the container ships with focus on the assessment of the influence of the wave-induced high-frequency hull girder vibrations. In several cases, the high-frequency contribution to the fatigue damage is dominating the estimated fatigue damage. Spectral formulations for estimating fatigue damage from bi-modal processes are explored and found to yield results fairly similar to the outcome of classical fatigue damage estimation from rainflow counting. However, in a few cases higher fatigue damage rates were estimated from rainflow counting than from narrow-band approximations.<br /><br />In summary and only considering larger container ships, the new and original contributions of the thesis are believed to be:<br /><br />•From full scale measurements the hull girder vibrational response is generally found to be dominated by the 2-node vertical bending mode even when the ship is sailing in oblique seas.<br />•The vertical bending moment in hogging and sagging is amplified considerably by the effect of the hull girder flexibility and the wave-induced hull girder vibrations are found capable of increasing the vertical wave bending moment amidships by 100% or more.<br />•The vertical hogging bending moment can be as critical as the sagging bending moment in design.<br />•From comparison of models tests and numerical methods, it seems that the numerical methods are not capable of fully capturing the effect of hull girder flexibility seen in model tests.<br />•The peak-over-threshold method is found to be the most useful method for extreme value prediction of the vertical bending moment in combination with an appropriate asymptotic extreme value distribution.</div><div class="phddissertation"><table class="properties"><tbody><tr class="language"><th>Original language</th><td>English</td></tr><tr><th>Publication date</th><td><span class="date">2014</span></td></tr></tbody></table><table class="properties"><tbody><tr><th>Place of publication</th><td>Kgs. Lyngby</td></tr><tr><th>Publisher</th><td><a class="link" rel="Publisher" href="http://orbit.dtu.dk/en/publishers/technical-university-of-denmark-dtu(f7c040ab-194a-4da4-b458-670da24b2a08).html"><span>Technical University of Denmark (DTU)</span></a></td></tr><tr><th>Number of pages</th><td>416</td></tr><tr><th>ISBN (print)</th><td> 978-87-7475-385-8</td></tr><tr class="status"><th>State</th><td>Published</td></tr></tbody></table></div> <table class="properties"><tbody><tr><th>Name</th><td>DCAMM Special Report</td></tr><tr><th>Number</th><td>S166</td></tr><tr><th>ISSN (print)</th><td>0903-1685</td></tr></tbody></table></div></div>Tue, 31 Dec 2013 23:00:00 GMThttp://orbit.dtu.dk/en/publications/full-scale-measurements-of-the-hydroelastic-response-of-large-container-ships-for-decision-support(80015fe7-b181-4e59-97fd-61f330063fe3).html2013-12-31T23:00:00ZEconomic Model Predictive Control for Large-Scale and Distributed Energy Systems
http://orbit.dtu.dk/en/publications/economic-model-predictive-control-for-largescale-and-distributed-energy-systems(8e96cd7e-66ed-4210-ba0c-36aa4ae36984).html
<div style='font-size: 9px;'><div class="rendering rendering_publication rendering_publication_short rendering_bookanthology rendering_short rendering_bookanthology_short"><h2 class="title"><a class="link" rel="BookAnthology" href="http://orbit.dtu.dk/en/publications/economic-model-predictive-control-for-largescale-and-distributed-energy-systems(8e96cd7e-66ed-4210-ba0c-36aa4ae36984).html"><span>Economic Model Predictive Control for Large-Scale and Distributed Energy Systems</span></a></h2><a class="link person" rel="Person" href="http://orbit.dtu.dk/en/persons/laura-standardi(f6f6929c-e6b2-4db7-aa06-0520df6b84c3).html"><span>Standardi, L.</span></a>, <a class="link person" rel="Person" href="http://orbit.dtu.dk/en/persons/john-bagterp-joergensen(d2bc6de4-a5c6-43f7-8712-606bd02a88e9).html"><span>Jørgensen, J. B.</span></a> & <a class="link person" rel="Person" href="http://orbit.dtu.dk/en/persons/niels-kjoelstad-poulsen(6c1bb106-d448-47cb-978c-3d9efee3ca97).html"><span>Poulsen, N. K.</span></a> <span class="date">2014</span> Kgs. Lyngby: <a class="link" rel="Publisher" href="http://orbit.dtu.dk/en/publishers/technical-university-of-denmark-dtu(f7c040ab-194a-4da4-b458-670da24b2a08).html"><span>Technical University of Denmark (DTU)</span></a>. <span class="numberofpages">179 p.</span> (DTU Compute PHD-2014; No. 356).<p class="type"><span class="type_family">Publication<span class="type_family_sep">: </span></span><span class="type_classification_parent">Research<span class="type_parent_sep"> › </span></span><span class="type_classification">Ph.D. thesis – Annual report year: 2014</span></p></div><div class="rendering rendering_publication rendering_publication_detailsportal rendering_bookanthology rendering_detailsportal rendering_bookanthology_detailsportal"><div class="textblock"><p class="MsoBodyText">In this thesis, we consider control strategies for large and distributed energy systems that are important for the implementation of smart grid technologies. An electrical grid has to ensure reliability and avoid long-term interruptions in the power supply. Moreover, the share of Renewable Energy Sources (RESs) in the smart grids is increasing. These energy sources bring uncertainty to the production due to their fluctuations. Hence,smart grids need suitable control systems that are able to continuously balance power production and consumption. We apply the Economic Model Predictive Control (EMPC) strategy to optimise the economic performances of the energy systems and to balance the power production and consumption. In the case of large-scale energy systems, the electrical grid connects a high number of power units. Because of this, the related control problem involves a high number of variables and constraints and its solution requires high computational times. Energy systems have a hierarchical control framework and the controllers have to work in the time-scale required by their hierarchy level. Dedicated optimisation techniques efficiently solve the control problem and reduce computational time. We implement the Dantzig-Wolfe decomposition technique to efficiently solve the EMPC problem.</p><p class="MsoBodyText">The contributions of this thesis are primarily on:</p><b>Large-scale energy system</b><p class="MsoBodyText">Smart-grids connect a high number of energy units. In such a large-scale scenario the energy units are independent and dynamically decoupled. The mathematical model of the large-scale energy system embodies the decoupled dynamics of each power units. Moreover,all units of the grid contribute to the overall power production.</p><p class="MsoBodyText"><b style="mso-bidi-font-weight:normal">Economic Model Predictive Control (EMPC)</b></p><p class="MsoBodyText">This control strategy is an extension of the Model Predictive Control (MPC)strategy. Energy systems often involve stochastic variables due to the share of fluctuating Renewable Energy Sources (RESs). Moreover, the related control problems are multi variables and they are hard, or impossible, to split into single-input-single-output control systems. MPC strategy can handle multi variables control problems and it can embody stochastic variables. The Economic MPC (EMPC) policy optimises the economic performances of the process. In this work, we apply the EMPC to energy systems and it computes the control trajectory for each energy unit. This control policy minimises production costs and ensures that the power production satisfies the customers’ demand. The EMPC designs a linear control problem that has a block-angular constraints matrix and it has two sets of constraints. The independent dynamics of the energy units define the decoupling constraints sited on the diagonal. The coupling constraints represent the common goal of all power units in the energy system and this is to satisfy the customers’ demand. The Dantzig-Wolfe optimisation technique applies to this structure of the constraints matrix in the view of fastening the control algorithm and increase its applicability.</p><b>Dantzig-Wolfe decomposition</b><p class="MsoBodyText">The Dantzig-Wolfe decomposition solves the EMPC problem through a distributed optimisation technique. The EMPC problem via Dantzig-Wolfe decomposition algorithm computes the optimal input trajectory for each energy unit and reduces the computation times. Moreover, such a control algorithm applies to large-scale energy systems and the number of energy units does not affect the performances of the controller. In this thesis, we also investigate suboptimal solutions of the EMPC problem via modified versions of the Dantzig-Wolfe decomposition algorithms. The feasibility of the suboptimal solutions suffices for stability. The goal of these modified Dantzig-Wolfe decomposition algorithms is to reduce computation time in the solution of the EMPC problem.</p></div><div class="phddissertation"><table class="properties"><tbody><tr class="language"><th>Original language</th><td>English</td></tr><tr><th>Publication date</th><td><span class="date">2014</span></td></tr></tbody></table><table class="properties"><tbody><tr><th>Place of publication</th><td>Kgs. Lyngby</td></tr><tr><th>Publisher</th><td><a class="link" rel="Publisher" href="http://orbit.dtu.dk/en/publishers/technical-university-of-denmark-dtu(f7c040ab-194a-4da4-b458-670da24b2a08).html"><span>Technical University of Denmark (DTU)</span></a></td></tr><tr><th>Number of pages</th><td>179</td></tr><tr class="status"><th>State</th><td>Submitted</td></tr></tbody></table></div> <table class="properties"><tbody><tr><th>Name</th><td>DTU Compute PHD-2014</td></tr><tr><th>Number</th><td>356</td></tr><tr><th>ISSN (print)</th><td>0909-3192</td></tr></tbody></table></div></div>Tue, 31 Dec 2013 23:00:00 GMThttp://orbit.dtu.dk/en/publications/economic-model-predictive-control-for-largescale-and-distributed-energy-systems(8e96cd7e-66ed-4210-ba0c-36aa4ae36984).html2013-12-31T23:00:00ZStatistical learning for predictive targeting in online advertising
http://orbit.dtu.dk/en/publications/statistical-learning-for-predictive-targeting-in-online-advertising(01a99509-9ab8-4c2c-b370-2d7ee4a873ea).html
<div style='font-size: 9px;'><div class="rendering rendering_publication rendering_publication_short rendering_bookanthology rendering_short rendering_bookanthology_short"><h2 class="title"><a class="link" rel="BookAnthology" href="http://orbit.dtu.dk/en/publications/statistical-learning-for-predictive-targeting-in-online-advertising(01a99509-9ab8-4c2c-b370-2d7ee4a873ea).html"><span>Statistical learning for predictive targeting in online advertising</span></a></h2><a class="link person" rel="Person" href="http://orbit.dtu.dk/en/persons/bjarne-oerum-fruergaard(b7bb8d93-b296-422a-8953-2483a0345537).html"><span>Fruergaard, B. Ø.</span></a>, <a class="link person" rel="Person" href="http://orbit.dtu.dk/en/persons/lars-kai-hansen(3bdd1c4b-6c83-462b-a4cd-1ffbdcbff128).html"><span>Hansen, L. K.</span></a> & <a class="link person" rel="Person" href="http://orbit.dtu.dk/en/persons/jesper-urban(0047bd4a-77f5-49c0-aea4-79ef1d1ef3c5).html"><span>Urban, J.</span></a> <span class="date">2014</span> Kgs. Lyngby: <a class="link" rel="Publisher" href="http://orbit.dtu.dk/en/publishers/technical-university-of-denmark-dtu(f7c040ab-194a-4da4-b458-670da24b2a08).html"><span>Technical University of Denmark (DTU)</span></a>. <span class="numberofpages">144 p.</span> (DTU Compute PHD-2014; No. 355).<p class="type"><span class="type_family">Publication<span class="type_family_sep">: </span></span><span class="type_classification_parent">Research<span class="type_parent_sep"> › </span></span><span class="type_classification">Ph.D. thesis – Annual report year: 2014</span></p></div><div class="rendering rendering_publication rendering_publication_detailsportal rendering_bookanthology rendering_detailsportal rendering_bookanthology_detailsportal"><div class="textblock">The focus in this thesis is investigation of machine learning methods with applications in computational advertising. Computational advertising is the broad discipline of building systems which can reach audiences browsing the Internet with targeted advertisements. At the core of such systems, algorithms are needed for making decisions. It is in one such particular instance of computational advertising, namely in web banner advertising, that we investigate machine learning methods to assist and make decisions in order to optimize the placements of ads. <br />The industrial partner in this work is Adform, an international online advertising technology partner. This also means that the analyses and methods in this work are developed with particular use-cases within Adform in mind and thus need also to be applicable in Adform’s technology stack. This implies extra thought on scalability and performance.<br />The particular use-case which is used as a benchmark for our results, is clickthrough rate prediction. In this task one aims to predict the probability that a user will click on an advertisement, based on attributes about the user, the advertisement the context, and other signals, such as time. This has its main application in real-time bidding ad exchanges, where each advertiser is given a chance to place bids for showing their ad while the page loads, and the winning bid gets to display their banner. <br />The contributions of this thesis entail application of a hybrid model of explicit and latent features for learning probabilities of clicks, which is a methodological extension of the current model in production at Adform. Our findings confirm that latent features can increase predictive performance in the setup of click-through rate prediction. They also reveal a tedious process for tuning the model for optimal performance.<br />We also present variations of Bayesian generative models for stochastic blockmodeling for inference of structure based on browsing patterns. Applying this structural information to improve click-through rate prediction becomes a two-step procedure; 1) learn user and URL profiles from browsing patterns, 2) use the profiles as additional features in a click-through rate prediction model. The assumption we implicitly make is reasonable: Users and URLs that are grouped together based on browsing patterns will have similar responses to ads, e.g., can be used as predictors of clicks. We report successful examples of applying this approach in practice. <br />Finally, we introduce the multiple-networks stochastic blockmodel (MNSBM), a model for efficient overlapping community detection in complex networks which can be assumed to be an aggregation of multiple block-structured subnetworks.<br /></div><div class="phddissertation"><table class="properties"><tbody><tr class="language"><th>Original language</th><td>English</td></tr><tr><th>Publication date</th><td><span class="date">2014</span></td></tr></tbody></table><table class="properties"><tbody><tr><th>Place of publication</th><td>Kgs. Lyngby</td></tr><tr><th>Publisher</th><td><a class="link" rel="Publisher" href="http://orbit.dtu.dk/en/publishers/technical-university-of-denmark-dtu(f7c040ab-194a-4da4-b458-670da24b2a08).html"><span>Technical University of Denmark (DTU)</span></a></td></tr><tr><th>Number of pages</th><td>144</td></tr><tr class="status"><th>State</th><td>Submitted</td></tr></tbody></table></div> <table class="properties"><tbody><tr><th>Name</th><td>DTU Compute PHD-2014</td></tr><tr><th>Number</th><td>355</td></tr><tr><th>ISSN (print)</th><td>0909-3192</td></tr></tbody></table></div></div>Tue, 31 Dec 2013 23:00:00 GMThttp://orbit.dtu.dk/en/publications/statistical-learning-for-predictive-targeting-in-online-advertising(01a99509-9ab8-4c2c-b370-2d7ee4a873ea).html2013-12-31T23:00:00ZNMR structural studies of oligosaccharides and other natural products
http://orbit.dtu.dk/en/publications/nmr-structural-studies-of-oligosaccharides-and-other-natural-products(7420df81-3623-497b-825c-09fa6e29c7db).html
<div style='font-size: 9px;'><div class="rendering rendering_publication rendering_publication_short rendering_bookanthology rendering_short rendering_bookanthology_short"><h2 class="title"><a class="link" rel="BookAnthology" href="http://orbit.dtu.dk/en/publications/nmr-structural-studies-of-oligosaccharides-and-other-natural-products(7420df81-3623-497b-825c-09fa6e29c7db).html"><span>NMR structural studies of oligosaccharides and other natural products</span></a></h2><a class="link person" rel="Person" href="http://orbit.dtu.dk/en/persons/louise-kjaerulff(840d3acf-77d5-4323-84ed-14a6ff8aabfc).html"><span>Kjærulff, L.</span></a> & <a class="link person" rel="Person" href="http://orbit.dtu.dk/en/persons/charlotte-held-gotfredsen(1cf4df84-2d0e-4693-bf47-e9932e3bf0e3).html"><span>Gotfredsen, C. H.</span></a> <span class="date">2014</span> <a class="link" rel="Publisher" href="http://orbit.dtu.dk/en/publishers/technical-university-of-denmark-dtu(f7c040ab-194a-4da4-b458-670da24b2a08).html"><span>Technical University of Denmark (DTU)</span></a>. <span class="numberofpages">311 p.</span><p class="type"><span class="type_family">Publication<span class="type_family_sep">: </span></span><span class="type_classification_parent">Research<span class="type_parent_sep"> › </span></span><span class="type_classification">Ph.D. thesis – Annual report year: 2014</span></p></div><div class="rendering rendering_publication rendering_publication_detailsportal rendering_bookanthology rendering_detailsportal rendering_bookanthology_detailsportal"><div class="textblock">NMR spectroscopy is an important tool in chemical analysis and for structural studies in various research areas. The subject of this thesis is liquid state NMR spectroscopy for structural analysis of small molecules, both regarding method development and structure elucidation by NMR spectroscopy.<br />HMBC+ is a new pseudo-3D NMR experiment for measurement of long-range homonuclear JHH coupling constants in small molecules. Based on two existing experiments, multiplicity edited HMBC and HAT HMBC by Benie, Nyberg, and Sørensen, the new HMBC+ experiment takes advantage of a π phase difference with respect to n+1JHH between these two experiments, observed in the nJCH HMBC cross peak. Through a double editing procedure this enables straightforward determination of both sign and magnitude of n+1JHH, including for very small coupling constants. Excellent results were obtained for the natural product strychnine, and 28 n+1JHH coupling constants were determined, of which 10 had not previously been measured due to their small sizes. By comparing to calculated coupling constants, an RMSD of 0.28 Hz was obtained, all coupling constant signs fit with the calculations, and all errors were below 1 Hz, also when comparing the absolute values to experimental coupling constants measured by Carter et al. using 1D 1H NMR on deuterated strychnine analogues. Providing an excellent tool for measurement of long-range JHH with the extra convenience of labeling the coupling constant information to the 13C chemical shift through the nJCH correlation, this experiment has exciting applications for configurational assignment of e.g. carbohydrates and for residual dipolar couplings.<br />Identification of known molecules and discovery of novel molecules are other important applications of NMR spectroscopy. Bacteria and fungi produce secondary metabolites for signaling and competing against other organisms, and these molecules are important in drug discovery due to their inherent biological activities. From a marine Photobacterium (P. halotolerans) we isolated the solonamides and the ngercheumicins, two families of cyclic depsipeptides capable of attenuating virulence in S. aureus by quorum sensing inhibition through the accessory gene regulator agr. This is likely due to their structural similarities with the auto-inducing peptides of S. aureus, controlling quorum sensing in bacterial populations. A filamentous fungus, A. fijiensis, was also investigated for production of novel secondary metabolites, and a new pyranonigrin (E) was isolated and structure elucidated by NMR spectroscopy along with JBIR-74 and decumbenone A, two known metabolites previously isolated from Aspergillus and Penicillium species.<br />Oligosaccharides found in human milk are important for infant nutrition, and a collaborative effort of university and industry partners was aimed at establishing methods for production of human milk oligosaccharides. Two different bioenzymatic methods for production of 3’-sialyllactose were investigated, and a screening of trans-fucosidases enabled the NMR spectroscopic identification of three pNP-fucosylfucopyranosides as major reaction products.</div><div class="phddissertation"><table class="properties"><tbody><tr class="language"><th>Original language</th><td>English</td></tr><tr><th>Publication date</th><td><span class="date">2014</span></td></tr></tbody></table><table class="properties"><tbody><tr><th>Publisher</th><td><a class="link" rel="Publisher" href="http://orbit.dtu.dk/en/publishers/technical-university-of-denmark-dtu(f7c040ab-194a-4da4-b458-670da24b2a08).html"><span>Technical University of Denmark (DTU)</span></a></td></tr><tr><th>Number of pages</th><td>311</td></tr><tr class="status"><th>State</th><td>Published</td></tr></tbody></table></div></div></div>Tue, 31 Dec 2013 23:00:00 GMThttp://orbit.dtu.dk/en/publications/nmr-structural-studies-of-oligosaccharides-and-other-natural-products(7420df81-3623-497b-825c-09fa6e29c7db).html2013-12-31T23:00:00ZAvailability by Design - A Complementary Approach to Denial-of-Service
http://orbit.dtu.dk/en/publications/availability-by-design--a-complementary-approach-to-denialofservice(d1ea8614-836f-4f8e-a6e8-f594f316aaab).html
<div style='font-size: 9px;'><div class="rendering rendering_publication rendering_publication_short rendering_bookanthology rendering_short rendering_bookanthology_short"><h2 class="title"><a class="link" rel="BookAnthology" href="http://orbit.dtu.dk/en/publications/availability-by-design--a-complementary-approach-to-denialofservice(d1ea8614-836f-4f8e-a6e8-f594f316aaab).html"><span>Availability by Design - A Complementary Approach to Denial-of-Service</span></a></h2><a class="link person" rel="Person" href="http://orbit.dtu.dk/en/persons/roberto-vigo(30a191fc-d2a3-4401-a0d9-41c4ab54de7c).html"><span>Vigo, R.</span></a>, <a class="link person" rel="Person" href="http://orbit.dtu.dk/en/persons/hanne-riis-nielson(44a7d08e-b083-4419-a466-8f4392a7112f).html"><span>Nielson, H. R.</span></a> & <a class="link person" rel="Person" href="http://orbit.dtu.dk/en/persons/flemming-nielson(f63ba194-185f-465d-9b2b-1f6ff489bc3d).html"><span>Nielson, F.</span></a> <span class="date">2014</span> Kgs. Lyngby: <a class="link" rel="Publisher" href="http://orbit.dtu.dk/en/publishers/technical-university-of-denmark-dtu(f7c040ab-194a-4da4-b458-670da24b2a08).html"><span>Technical University of Denmark (DTU)</span></a>. <span class="numberofpages">190 p.</span> (DTU Compute PHD-2014; No. 353).<p class="type"><span class="type_family">Publication<span class="type_family_sep">: </span></span><span class="type_classification_parent">Research<span class="type_parent_sep"> › </span></span><span class="type_classification">Ph.D. thesis – Annual report year: 2014</span></p></div><div class="rendering rendering_publication rendering_publication_detailsportal rendering_bookanthology rendering_detailsportal rendering_bookanthology_detailsportal"><div class="textblock">In computer security, a Denial-of-Service (DoS) attack aims at making a resource unavailable. DoS attacks to systems of public concern occur increasingly and have become infamous on the Internet, where they have targeted major corporations and institutions, thus reaching the general public. There exist various practical techniques to face DoS attacks and mitigate their effects, yet we witness the successfulness of many.<br />The need for a renewed investigation of availability gains in relevance when considering that our life is more and more dominated by Cyber-Physical Systems (CPSs), large-scale network of sensors that interact with the physical environment. CPSs are increasingly exploited in the realisation of critical infrastructure, from the power grid to healthcare, traffic control, and defence applications. Such systems are particularly prone to DoS attacks: in addition to classic communication-based attacks, their components can be subject to physical capture. Moreover, sensors are often powered by batteries, and time-limited unavailability is usually a stage planned to prolong their life span.<br />This dissertation argues that techniques rooted in the theory and practice of programming languages, language-based techniques, offer a unifying framework to deal with the consequences of DoS, thereby encompassing inadvertent and malicious sources of unavailability in a uniform manner.<br />In support to this claim we develop a family of process calculi, the Quality Calculi, where availability considerations are promoted to be first-class object of the language domain. Moreover, these modelling tools are complemented by static analyses that pinpoint where and why unavailability may occur, levering the enhanced expressiveness of the language.<br />The ultimate aim of the framework is to foster the development of systems resilient to DoS by means of a principled design process, in which formal models allow, and verification tools enforce, the production of such robust code.<br /></div><div class="phddissertation"><table class="properties"><tbody><tr class="language"><th>Original language</th><td>English</td></tr><tr><th>Publication date</th><td><span class="date">2014</span></td></tr></tbody></table><table class="properties"><tbody><tr><th>Place of publication</th><td>Kgs. Lyngby</td></tr><tr><th>Publisher</th><td><a class="link" rel="Publisher" href="http://orbit.dtu.dk/en/publishers/technical-university-of-denmark-dtu(f7c040ab-194a-4da4-b458-670da24b2a08).html"><span>Technical University of Denmark (DTU)</span></a></td></tr><tr><th>Number of pages</th><td>190</td></tr><tr class="status"><th>State</th><td>Submitted</td></tr></tbody></table></div> <table class="properties"><tbody><tr><th>Name</th><td>DTU Compute PHD-2014</td></tr><tr><th>Number</th><td>353</td></tr><tr><th>ISSN (print)</th><td>0909-3192</td></tr></tbody></table></div></div>Tue, 31 Dec 2013 23:00:00 GMThttp://orbit.dtu.dk/en/publications/availability-by-design--a-complementary-approach-to-denialofservice(d1ea8614-836f-4f8e-a6e8-f594f316aaab).html2013-12-31T23:00:00ZCUDArray: CUDA-based NumPy
http://orbit.dtu.dk/en/publications/cudarray-cudabased-numpy(1e806793-a31e-4a87-bc14-7d389f0a2d57).html
<div style='font-size: 9px;'><div class="rendering rendering_publication rendering_publication_short rendering_bookanthology rendering_short rendering_bookanthology_short"><h2 class="title"><a class="link" rel="BookAnthology" href="http://orbit.dtu.dk/en/publications/cudarray-cudabased-numpy(1e806793-a31e-4a87-bc14-7d389f0a2d57).html"><span>CUDArray: CUDA-based NumPy</span></a></h2><a class="link person" rel="Person" href="http://orbit.dtu.dk/en/persons/anders-boesen-lindbo-larsen(e16ce50c-66b8-488b-b2bd-36c64f111f76).html"><span>Larsen, A. B. L.</span></a> <span class="date">2014</span> Kgs. Lyngby: <a class="link" rel="Publisher" href="http://orbit.dtu.dk/en/publishers/technical-university-of-denmark-dtu(f7c040ab-194a-4da4-b458-670da24b2a08).html"><span>Technical University of Denmark (DTU)</span></a>. <span class="numberofpages">5 p.</span> (DTU Compute-Technical Report-2014; No. 21).<p class="type"><span class="type_family">Publication<span class="type_family_sep">: </span></span><span class="type_classification_parent">Research<span class="type_parent_sep"> › </span></span><span class="type_classification">Report – Annual report year: 2014</span></p></div><div class="rendering rendering_publication rendering_publication_detailsportal rendering_bookanthology rendering_detailsportal rendering_bookanthology_detailsportal"><div class="textblock">This technical report introduces CUDArray – a CUDA-accelerated subset of the NumPy library. The goal of CUDArray is to combine the ease of development from NumPy with the computational power of Nvidia GPUs in a lightweight and extensible framework. Since the motivation behind CUDArray is to facilitate neural network programming, CUDArray extends NumPy with a neural network submodule. This module has both a CPU and a GPU back-end to allow for experiments without requiring a GPU.</div><div class="scientificreport"><table class="properties"><tbody><tr class="language"><th>Original language</th><td>English</td></tr><tr><th>Publication date</th><td><span class="date">2014</span></td></tr></tbody></table><table class="properties"><tbody><tr><th>Place of publication</th><td>Kgs. Lyngby</td></tr><tr><th>Publisher</th><td><a class="link" rel="Publisher" href="http://orbit.dtu.dk/en/publishers/technical-university-of-denmark-dtu(f7c040ab-194a-4da4-b458-670da24b2a08).html"><span>Technical University of Denmark (DTU)</span></a></td></tr><tr><th>Number of pages</th><td>5</td></tr><tr class="status"><th>State</th><td>Published</td></tr></tbody></table></div> <table class="properties"><tbody><tr><th>Name</th><td>DTU Compute-Technical Report-2014</td></tr><tr><th>Number</th><td>21</td></tr><tr><th>ISSN (print)</th><td>1601-2321</td></tr></tbody></table></div></div>Tue, 31 Dec 2013 23:00:00 GMThttp://orbit.dtu.dk/en/publications/cudarray-cudabased-numpy(1e806793-a31e-4a87-bc14-7d389f0a2d57).html2013-12-31T23:00:00ZAn algebraic approach to graph codes
http://orbit.dtu.dk/en/publications/an-algebraic-approach-to-graph-codes(bb4e0715-71b9-406f-bf7e-84ccf1441b87).html
<div style='font-size: 9px;'><div class="rendering rendering_publication rendering_publication_short rendering_bookanthology rendering_short rendering_bookanthology_short"><h2 class="title"><a class="link" rel="BookAnthology" href="http://orbit.dtu.dk/en/publications/an-algebraic-approach-to-graph-codes(bb4e0715-71b9-406f-bf7e-84ccf1441b87).html"><span>An algebraic approach to graph codes</span></a></h2><a class="link person" rel="Person" href="http://orbit.dtu.dk/en/persons/fernando-pinero(3f66e5f4-861e-4384-89ad-fddfc12127a2).html"><span>Pinero, F.</span></a> & <a class="link person" rel="Person" href="http://orbit.dtu.dk/en/persons/peter-beelen(da3e5b30-4aed-4f4f-9e17-3fda35fb6788).html"><span>Beelen, P.</span></a> <span class="date">2014</span> Kgs. Lyngby: <a class="link" rel="Publisher" href="http://orbit.dtu.dk/en/publishers/technical-university-of-denmark-dtu(f7c040ab-194a-4da4-b458-670da24b2a08).html"><span>Technical University of Denmark (DTU)</span></a>. <span class="numberofpages">94 p.</span> (DTU Compute PHD-2014; No. 352).<p class="type"><span class="type_family">Publication<span class="type_family_sep">: </span></span><span class="type_classification_parent">Research<span class="type_parent_sep"> › </span></span><span class="type_classification">Ph.D. thesis – Annual report year: 2014</span></p></div><div class="rendering rendering_publication rendering_publication_detailsportal rendering_bookanthology rendering_detailsportal rendering_bookanthology_detailsportal"><div class="textblock">This thesis consists of six chapters. The first chapter, contains a short introduction to coding theory in which we explain the coding theory concepts we use. In the second chapter, we present the required theory for evaluation codes and also give an example of some fundamental codes in coding theory as evaluation codes. Chapter three consists of the introduction to graph based codes, such as Tanner codes and graph codes. In Chapter four, we compute the dimension of some graph based codes with a result combining graph based codes and subfield subcodes. Moreover, some codes in chapter four are optimal or best known for their parameters. In chapter five we study some graph codes with Reed–Solomon component codes. The underlying graph is well known and widely used for its good characteristics. This helps us to compute the dimension of the graph codes. We also introduce a combinatorial concept related to the iterative encoding of graph codes with MDS component code. The last chapter deals with affine Grassmann codes and Grassmann codes. We begin with some previously known codes and prove that they are also Tanner codes of the incidence graph of the point–line partial geometry of the Grassmannian. We expect that the techniques exposed in chapter six are also applicable to other codes as well.</div><div class="phddissertation"><table class="properties"><tbody><tr class="language"><th>Original language</th><td>English</td></tr><tr><th>Publication date</th><td><span class="date">2014</span></td></tr></tbody></table><table class="properties"><tbody><tr><th>Place of publication</th><td>Kgs. Lyngby</td></tr><tr><th>Publisher</th><td><a class="link" rel="Publisher" href="http://orbit.dtu.dk/en/publishers/technical-university-of-denmark-dtu(f7c040ab-194a-4da4-b458-670da24b2a08).html"><span>Technical University of Denmark (DTU)</span></a></td></tr><tr><th>Number of pages</th><td>94</td></tr><tr class="status"><th>State</th><td>Submitted</td></tr></tbody></table></div> <table class="properties"><tbody><tr><th>Name</th><td>DTU Compute PHD-2014</td></tr><tr><th>Number</th><td>352</td></tr><tr><th>ISSN (print)</th><td>0909-3192</td></tr></tbody></table></div></div>Tue, 31 Dec 2013 23:00:00 GMThttp://orbit.dtu.dk/en/publications/an-algebraic-approach-to-graph-codes(bb4e0715-71b9-406f-bf7e-84ccf1441b87).html2013-12-31T23:00:00ZOn the use of functional calculus for phase-type and related distributions
http://orbit.dtu.dk/en/publications/on-the-use-of-functional-calculus-for-phasetype-and-related-distributions(6a1a9683-151c-46f9-9f89-d5ca1ac9ca47).html
<div style='font-size: 9px;'><div class="rendering rendering_publication rendering_publication_short rendering_bookanthology rendering_short rendering_bookanthology_short"><h2 class="title"><a class="link" rel="BookAnthology" href="http://orbit.dtu.dk/en/publications/on-the-use-of-functional-calculus-for-phasetype-and-related-distributions(6a1a9683-151c-46f9-9f89-d5ca1ac9ca47).html"><span>On the use of functional calculus for phase-type and related distributions</span></a></h2>Bladt, M., Campillo Navarro, A. & <a class="link person" rel="Person" href="http://orbit.dtu.dk/en/persons/bo-friis-nielsen(364b98b5-e608-4186-a566-62fa91a37d59).html"><span>Nielsen, B. F.</span></a> <span class="date">2014</span> Kgs. Lyngby: <a class="link" rel="Publisher" href="http://orbit.dtu.dk/en/publishers/technical-university-of-denmark-dtu(f7c040ab-194a-4da4-b458-670da24b2a08).html"><span>Technical University of Denmark (DTU)</span></a>. <span class="numberofpages">25 p.</span> (DTU Compute-Technical Report-2014; No. 18).<p class="type"><span class="type_family">Publication<span class="type_family_sep">: </span></span><span class="type_classification_parent">Research<span class="type_parent_sep"> › </span></span><span class="type_classification">Report – Annual report year: 2014</span></p></div><div class="rendering rendering_publication rendering_publication_detailsportal rendering_bookanthology rendering_detailsportal rendering_bookanthology_detailsportal"><div class="textblock">The area of phase-type distributions is renowned for its ability to obtain closed form formulas or algorithmically exact solutions to many complex stochastic models. The method of functional calculus will provide an additional tool along these lines for establishing results in terms of functions of matrices. Functional calculus, which is a branch of operator theory frequently associated with complex analysis, can be applied to phase-type and matrix-exponential distributions in a rather straightforward way. In this paper we provide a number of examples on how to execute the formal arguments.<br /></div><div class="scientificreport"><table class="properties"><tbody><tr class="language"><th>Original language</th><td>English</td></tr><tr><th>Publication date</th><td><span class="date">2014</span></td></tr></tbody></table><table class="properties"><tbody><tr><th>Place of publication</th><td>Kgs. Lyngby</td></tr><tr><th>Publisher</th><td><a class="link" rel="Publisher" href="http://orbit.dtu.dk/en/publishers/technical-university-of-denmark-dtu(f7c040ab-194a-4da4-b458-670da24b2a08).html"><span>Technical University of Denmark (DTU)</span></a></td></tr><tr><th>Number of pages</th><td>25</td></tr><tr class="status"><th>State</th><td>Published</td></tr></tbody></table></div> <table class="properties"><tbody><tr><th>Name</th><td>DTU Compute-Technical Report-2014</td></tr><tr><th>Number</th><td>18</td></tr><tr><th>ISSN (print)</th><td>1601-2321</td></tr></tbody></table></div></div>Tue, 31 Dec 2013 23:00:00 GMThttp://orbit.dtu.dk/en/publications/on-the-use-of-functional-calculus-for-phasetype-and-related-distributions(6a1a9683-151c-46f9-9f89-d5ca1ac9ca47).html2013-12-31T23:00:00Z