No description available.

No description available.

No description available.

Les mécanismes d’assortiment, sont des marchés dont l’objet est de réaliser une allocation économique efficace mais qui opèrent sans échange monétaire. L’efficacité d’un mécanisme peut être évaluée de manière théorique, mais il est aussi important d’évaluer sa performance avec des agents réels pour tenir compte des biais comportementaux et leur rationalité limitée. La thèse résumée dans cette note s’inscrit dans cette démarche en fournissant des résultats empiriques qui permettront d’améliorer l’utilisation des mécanismes d’appariement sur le terrain. Le premier chapitre de la thèse vise à analyser l’expérience acquise par des agents ayant participé à un mécanisme d’acceptation différée qui peut être transmise à de nouveaux participants pour améliorer l’efficacité du mécanisme. Les résultats obtenus montrent que les sujets sont plus susceptibles de révéler leur préférence réelle dans leur proposition quand ils reçoivent des conseils de leurs pairs des sessions précédentes. Le deuxième chapitre de la thèse étudie l’effet de la taille du marché sur la performance de deux mécanismes d’appariement. Les résultats montrent que l’augmentation du marché de 4 à 40 joueurs accroit le taux des participants qui révèlent leur préférence réelle dans leur proposition au mécanisme d’acceptation différée, mais que ce taux diminue dans le mécanisme de Boston. Le passage à une taille de marché de 4000 joueurs n’a pas d’effet supplémentaire significatif. Le troisième chapitre de la thèse justifie l’évolution du système d’admission des Universités en Chine. Des études montrent que le mécanisme de Boston n’élimine pas l’envie justifiée, qu’il est manipulable, et qu’il n’est pas Pareto-efficace. Le mécanisme d’allocation simple élimine l’envie justifiée, et il est non-manipulable et Pareto-efficace. Ce résultat justifie la transition récente de l’algorithme de choix séquentiel à l’algorithme de choix parallèle dans les pratiques d’admission aux Universités chinoises. / This thesis consists of three essays examining empirical factors that are important for the success of the matching mechanism in the real world. The first chapter discusses whether highly experienced people can transmit what they have learned and encourage new participants to reveal their true preferences under the Deferred Acceptance mechanism. I address this issue in a laboratory experiment to check the effect of peer experience on individual behaviors and the performance of the DA mechanism. Results show that subjects are more likely to play truthfully when learning advice from their peers in previous sessions.The second chapter studies the performance of the Boston and the DA mechanism in a laboratory under different scale of the matching markets. Results show that increasing the market size from 4 to 40 students per match increases participant truth-telling under the DA but decrease it under the Boston mechanism, leading to a decrease in efficiency but no change in the large stability advantage of the DA over the Boston mechanism. However, further increase in the scale to 4,000 students per match has no effect.The third chapter justifies the evolution of the college admissions system in China from a mechanism design perspective. Studies show that the Boston mechanism does not eliminate justified envy, is not strategy-proof and is not Pareto-efficient. The Simple Serial Dictatorship mechanism eliminates justified envy, is strategy-proof and is Pareto-efficient, thus outperforming the Boston mechanism in all three criteria. This result provides justification for the mechanism transition in recent years in China’s college admissions practices.

[Truncated abstract] Various capabilities such as the ability to read or conduct a conversation rely on our ability to maintain and recall information in the correct order. Research spanning more than a century has been devoted to understanding how units of information are retained in order in short-term memory. The nature of the mechanisms that code the positions of items in serial short-term verbal recall can be investigated by examining a set of phenomena that can be termed temporal grouping effects. Inserting extended pauses to break a list of verbal items into sub-lists (e.g. SHD-QNR-BJF, where the dashes represents the pauses) improves the accuracy of serial recall relative to performance observed without this temporal grouping. In addition, two other effects are linked to temporal grouping. One of these effects is a shift in the shape of the serial position function, which changes from a single bowed function to a multiple-bowed function. That is, the serial position curve for ungrouped sequences is typically characterized by better performance for the beginning and ending items compared to the mid-list items. For grouped lists, the multiple-bowed function comprises better recall for the beginning and ending items within each group. Another effect associated with temporal grouping is a change in the patterns of order errors. For ungrouped sequences (e.g. SHDQNRBJF), order errors often involve the swapping of items in neighbouring positions, such as exchanging D for Q or R for B. By contrast, grouped sequences (such as SHD-QNR-BJF) show a reduction in order errors that cross group boundaries such as exchanging items D and Q or R and B; instead, there tend to be an increased incidence of exchanging items that share similar within-group positions such as swapping H and N or Q and B. According to several current models of short-term memory, items are retained by associating them with extra-list information such as contextual information. ... This was done by unconfounding temporal position (time from group onset) and ordinal position (number of items from group onset) for certain key items in sequences comprising two groups of four consonants. The critical manipulation was to vary the SOAs within and across the two groups. Errors that involve items migrating across groups should preserve within-group temporal position according to oscillator models, but should preserve within-group ordinal position according to non-oscillator models. Results from the intergroup errors strongly favored preservation of ordinal rather than temporal position. Finally, the Appendix reports an unpublished experiment that examined patterns of errors in recalling sequences of nine visually presented letters, where the letters were grouped into threes using temporal gaps. A critical manipulation was the insertion of a tobe- ignored item (an asterisk) between the first and second letters of selected groups. Inclusion of this item failed to alter the patterns of errors observed, indicating that the coding of serial position is based on only those events represented for recall. The central conclusion based on all the studies is that serial order for verbal items is retained using contextual positional codes that change with each presentation of a tobe- remembered item, are influenced by large temporal gaps that lead to grouping, but otherwise are not dependent on the timing of events.

<p>Bit-serial multiplication with a fixed coefficient is commonly used in integrated circuits, such as digital filters and FFTs. These multiplications can be implemented using basic components such as adders, subtractors and D flip-flops. Multiplication with the same coefficient can be implemented in many ways, using different structures. Other studies in this area have focused on how to minimize the number of adders/subtractors, and often assumed that the cost for D flip-flops is neglectable. That simplification has been proved to be far too great, and further not at all necessary. In digital devices low power consumption is always desirable. How to attain this in bit-serial multipliers is a complex problem. </p><p>The aim of this thesis was to find a strategy on how to implement bit-serial multipliers with as low cost as possible. An important step was achieved by deriving formulas that can be used to calculate the carry switch probability in the adders/subtractors. It has also been established that it is possible to design a power model that can be applied to all possible structures of bit- serial multipliers.</p>

<p>I digitala kretsar kan både bit-parallella och bit-seriella interface förekomma.T ex kan en integrerad digital krets (IC-krets) internt använda sig av bit-parallella aritmetiska kretsar medan dess kommunikation med andra integrerade kretsar sker bit-seriellt. Genom att använda seriell kommunikation mellan IC-kretsar kan antalet ben på kapslarna effektivt begränsas.</p><p>Detta examensarbete gick ut på att göra en layoutgenerator för generering av en parametriserbar serie/parallellomvandlare och en parallell/serieomvandlare. När en krets använder sig av både bit-seriell och bit-parallell representation av data behövs dessa för att omvandla mellan formaten.</p> / <p>In digital circuits both bit-parallel and bit-serial interfaces can occur. For example, an integrated digital circuit can use bit-parallel arithmetic circuits internally while its communication with other integrated circuits is bit serial. By using serial communication between IC:s the number of pins on the packages can be effectively limited. </p><p>The purpose of this final project work was to create a layout generator for generation of a parametrizable serial/parallel converter and a parallel/serial converter. If a circuit uses both bit-serial and bit-parallel representation of data these are needed to convert between the formats.</p>

We examined relations between eye movements (single-fixation durations) and RSVP-based event-related potentials (ERPs; N400’s) recorded during reading the same sentences in two independent experiments. Longer fixation durations correlated with larger N400 amplitudes. Word frequency and predictability of the fixated word as well as the predictability of the upcoming word accounted for this covariance in a path-analytic model. Moreover, larger N400 amplitudes entailed longer fixation durations on the next word, a relation accounted for by word frequency. This pattern offers a neurophysiological correlate for the lag-word frequency effect on fixation durations: Word processing is reliably expressed not only in fixation durations on currently fixated words, but also in those on subsequently fixated words.

Bacterial chromosomes are highly dynamic, continuously changing with respect to gene content and size via a number of processes, including deletions that result in gene loss. How deletions form and at what rates has been the focus of this thesis. In paper II we investigated how chromosomal location affects chromosomal deletion rates in S. typhimurium. Deletion rates varied more than 100-fold between different chromosomal locations and some large deletions significantly increased the exponential growth rate of the cells. Our results suggest that the chromosome is heterogeneous with respect to deletion rates and that deletions may be genetically fixed as a consequence of natural selection rather than by drift or mutational biases. In paper I we examined in a laboratory setting how rapidly reductive evolution, i.e. gene loss, could occur. Using a serial passage approach, we showed that extensive genome reduction potentially could occur on a very short evolutionary time scale. For most deletions we observed little or no homology at the deletion endpoints, indicating that spontaneous deletions often form through a RecA independent process. In paper III we examined further how large spontaneous deletions form and, unexpectedly, showed that 90% of all spontaneous chromosomal deletions required error-prone translesion DNA polymerases for their formation. We propose that the translesion polymerases stimulate deletion formation by allowing extension of misaligned single-strand DNA ends. In paper IV we investigated how the translesion DNA polymerase Pol IV, RpoS and different types of stresses affect mutation rates in bacteria. Derepression of the LexA regulon caused a small to moderate increase in mutation rates that was fully dependent on functional endonucleases but only partly dependent on translesion DNA polymerases. RpoS levels and growth stresses had only minor effects on mutation rates. Thus, mutation rates appear very robust and are only weakly affected by growth conditions and induction of translesion polymerases and RpoS.

Swedish national statistics suggest that the number of reported stranger rapes is steadily increasing. Stranger rape is one of the most difficult types of crime for the police to investigate because there is no natural tie between the victim and offender. As a result, there is a need for more knowledge about how crime scene features could be used to make inferences of likely offender characteristics that could help investigators narrow down the pool of suspects. The aim in Study I was to examine how offender behaviors interact with contextual features, victim behaviors, and the assault outcome. Results suggest that the stranger rapes could be distinguished by five different dynamic rape pattern themes, which mainly differed on two dimensions: level of violence to control the victim, and level of impulsivity/premeditation characterizing the rapes. The results also highlight the importance of including contextual features when studying offender behaviors. The aim in Study II was to examine how single-victim rapists and serial rapists can be differentiated by the actions at their first stranger rape. Results suggest that three behaviors in conjunction: kissed victim, controlled victim, and offender drank alcohol before the offense, could be used to predict whether the offender was a single-victim rapist or serial rapist with a classification accuracy of 80.4 %. The aim in Study III was to examine how stranger rapists could be differentiated from a normative sample on background characteristics, and if stranger rapists’ pre-assault and initial-attack behaviors could be used to predict likely offender characteristics. Results showed that the strongest predictions could be made for previous criminal convictions, offender age, and the distance traveled by the offender to offend. Overall, the present thesis has found some scientific support for the use of crime scene behaviors to make inferences of likely offender characteristics that could be useful for profiling purposes. / <p>At the time of the doctoral defense, the following papers were unpublished and had a status as follows: Paper 1: Submitted. Paper 3: Submitted.</p>