Des approches statistiques-probabilites du comportement mécanique des composites à matrice céramique, fondées sur la description des phénomènes stochastiques à l'origine de la fragmentation matricielle et de la rupture ultime, ont été développées. Elles ont été appliquées à des composites unidirectionnels et tissés de type SiC/SiC. Elles ont permis de calculer le comportement mécanique et la rupture en traction et en flexion. L'approche de la fragmentation matricielle repose sur l'analyse des populations de défauts qui provoquent cette fragmentation. Dans le cas des composites unidirectionnels étudiés, une seule population est suffisante pour calculer le comportement. Pour les composites tissés étudiés, trois populations de défauts caractérisent le processus de fissuration. En ce qui concerne la rupture, une classification, fondée sur les effets de volumes et les ruptures successives d'éprouvettes en traction, permet de définir convenablement le choix du critère probabiliste de rupture (Weibull, Ergodique ...). Ainsi le caractère ergodique de la rupture d'un composite tissé SiC/SiC a été mis en évidence et vérifié sur des essais de flexion 3 points et 4 points.

Localization accuracy based on RSS (Received Signal Strength) is notoriously inaccurate in the application of wireless sensor networks. RSS is subject to shadowing effects, which is signal attenuation caused by stationary objects in the radio propagation. RSS are actually the result of decay over distances, and random attenuation over different directions. RSS measurement is also affected by antenna orientation. Starting from extracting the statistical orders in the function relationship between RSS and distance, this thesis first shows how non-metric MDS (Multi-Dimensional Scaling) is the suitable method for cooperative localization. Then, several issues are presented and discussed in the application of non-metric MDS, including determining full connections to avoid flip ambiguities, leveraging the proper initial estimation to avert from local minimum solutions, and imposing structural information to bend the localization result to a priori knowledge. To evaluate the solution, data were acquired from different scenarios including accurate radio propagation model, indoor empirical test, and outside empirical test. Experiment results shows that non-metric MDS can only combat the small scale randomness in the shadowing effects. To combat the large scale ones, macro-diversity approaches are further presented including rotating the receiver’s antenna or collecting RSS from more than one mote in the same position. By averaging the measurements from these diversified receivers, simulation results and empirical tests show that shadowing effects can be greatly reduced. Also included in this thesis is how effective packet structures should be designed in the mote programming based on TinyOS to collect different sequences of RSS measurements and fuse them together. / UOIT

This paper examines the possible contribution of evolutionary economics to macro-economic modelling of flood impacts to provide guidance for future economic risk modelling. Most macro-economic models start from a neoclassical economic perspective and focus on equilibrium outcomes, either in a static or dynamic way, and describe economic processes at a high level of aggregation. As a consequence, they typically fail to account for the complexity of social interactions and other behavioural responses of consumers and producers to disasters, which may affect the macroeconomic impacts of floods. Employing evolutionary principles and methods, such as agent-based modelling, may help to address some of the shortcomings of current macro-economic models. We explore and discuss the implications of applying consumer and producer heterogeneity, bounded rationality, network effects, social and technological learning, co-evolution and adaptive policy-making concepts into existing economic frameworks for the assessment of macro-economic impacts of floods. (authors' abstract)

With the prevalence of legacy C/C++ code, issues of readability and maintainability have become increasingly important. When we consider the problem of refactoring or migrating C/C++ code, we see the significant role that preprocessor directives play. It is partially because of these preprocessor directives that code maintenance has become extremely difficult. This thesis describes a method of fact extraction and code manipulation to create a set of transformations which will remove preprocessor directives from the original source, converting them into regular C/C++ code with as few changes as possible, while maintaining readability in the code. In addition, some of the subtle issues that may arise when migrating preprocessor directives are explored. After discussing the general architecture of the test implementation, an examination of some metrics gathered by running it on two software systems is given.

The purpose of this thesis is to study the current status and past trend of nursing turnover in Taiwan. The factors affecting nursing turnover are investigated. The research adopts longitudinal methodology and is based on the Medical Staff Database of the Department of Health, from which 420,183 samples between 1999 and 2005 are analyzed. The study results show that: 1. The registered nurses are more than non-registered ones. The average age and seniority of the nursing turnover personnel are 26 and 4.73, respectively. In hospitals, nursing turnover is higher in regional and district hospitals. In non-hospitals, the percentage of nursing turnover in home health care and nursing home is higher and increasing year by year. 2. The nursing turnover rate is closely related with economic growth rate. That is, the more the economic growth, the higher the nursing turnover. 3. With reference point in 1999, the growth rate of nursing turnover is negative from 2000 to 2001 and positive from 2002 to 2005, though a drop of 8.67% happened in 2005 when compared with 2004. Besides, the growth rate of nursing turnover in hospitals is inversely proportional to that in non-hospitals. 4. The policy of Global Budget Payment System has noticeable impact on nursing turnover. The phenomenon is most significant in clinics of the hospital organizations and home health care, nursing home and nursing school of the non-hospital organizations.

The purpose of this study is to assess the effectiveness of an intermediate textbook titled &ldquo / New Bridge to Success 3&rdquo / , which was prepared by Ministry of National Education as an instructional material for the ninth grade high school students from the perspectives of the teachers and students. The evaluation of the textbook concerned was conducted at macro level on the basis of eleven criteria. Both quantitative and qualitative data were obtained through student questionnaires administered to 336 students and interviews with eight teachers. Frequencies, percentages, means and standard deviations were calculated for each item to describe the overall picture of how the students rated the textbook in terms of ten criteria. Data collected through interviews were transcribed, content- analyzed and grouped according to 11 criteria used in this study for the evaluation of the particular textbook The results revealed that both teachers and students felt negative about the most of the characteristics of the textbook. It was found that the reading passages needed to be simplified in terms of both vocabulary load and structures. Majority of the students and all the teachers mentioned that the level of the textbook was not appropriate for the particular age group. It also indicated that the materials failed to consider learning style preferences of the visual, auditory, and kinesthetic learner. On the other hand, one of the strengths of the textbook was the artwork&rsquo / s being up-to date and helpful for the students to understand the lesson

The topic of insider/director trading raises interesting questions and has generated much attention from researchers, market participants and regulators. The purpose of this thesis is to investigate the long-run director trading performance in the UK market. It examines relationship between aggregate director trading and indicators of the UK macroeconomy focus on the macro-aspects in Chapter 4 and 5. The extant empirical literature on aggregate director trading can be categorized into two parts: the first is the relationship between director trading and the stock market; and the second is the link between stock returns and future aggregate economic activities. Having examined the macro-picture, it goes to examine a more micro-picture. Chapter 6 examines long-run relationship between director trading and market reactions. This thesis is organized around three research studies which are presented in Chapters 4, 5 and 6 and which examine long-run director trading activities in the UK. Chapters 4 and 5 together investigate the evidence for director trading activities and the macroeconomy. There is little literature on aggregate director trading and the macro-economy: therefore Chapter 4 examines the relationship between aggregate director trading and future market excess returns. Empirical evidence is presented which demonstrates that the returns on stock market are significantly correlated to future economic growth. Chapter 5 then examines whether the forecasting ability can be improved by adding aggregate director trading as a measurement of business confidence into the forecasting model. Chapter 6 examines the long-run performance of market reaction to director roles. In order to examine the relationship between aggregate director trading and the macro-economy, the link between aggregate director trading and future market excess returns is investigated. This thesis considers the importance of the seasonality issue in UK director trading and employs a number of alternative seasonality adjustments to adjust the raw data on aggregate director trading. The positive correlation between aggregate director trading and future market excess returns is confirmed and evidence is provided that indicates directors are contrarian: in aggregate they purchase (sell) their own-company stock prior to general stock market increases (decreases). In the long-run, the empirical work demonstrates that aggregate director trading has forecasting power in terms of predicting future stock market excess returns. Additional findings are that aggregate director trading in large firms has a positive significant predictive ability for identifying future excess returns of large firms and aggregate director trading of some industries has positive significant forecasting ability for future excess returns of these industries. Having confirmed the relationship between aggregate director trading and future market movement, this thesis turns to examine the link between aggregate director trading and future UK economic growth. It measures economic growth of future real economic activity by the change in gross domestic product (GDP) and it documents a strong correlation between past aggregate director trading and future real economic activity. The predictability of future economic growth increases with both the length of forecasting horizon and past net number of director trading. In a multivariate regression analysis this thesis finds that aggregate director trading retains predicting power with respect to future GDP growth even after including popular business cycle variables (dividend yield of FTSE All share, growth rate of industrial production and term spread) as explanatory variables. This finding suggests that aggregate director trading captures things related to changes in real activity but not captured by market factors (Fama-French 3 factors: SMB, HML and RMRF) and business cycle variables. After examining the relationship between aggregate director trading, market returns and changes in GDP, the last empirical Chapter of the thesis concentrates director trading on the micro-aspects of director trading and stock movement. It examines the stock market reaction to director trading with firm characteristics and the effects of director trading pattern. Using long-run calendar-time abnormal returns (CTAR) methodology with Fama-French 3-factor model, evidence is presented that directors do have more valuable information allowing them to make significant abnormal returns than other market participants, the performance of CFOs supports the information hierarchy hypothesis in 1- and 6-month post-purchase trading time, and the director trading with firm characteristics has a significant effect on stock abnormal returns.

In the context of regionalism and regionalisation, the notion of a region holds sway. Discussions around the topic of how spatial entities defined as ‘regions’ form, evolve, develop, become institutionalised and sometimes stabilized have been active and profound over the last forty years. Moreover, rich and diverse literature provides different conceptualizations and problematisation of regions that have been changing over time. ‘Heterogeneous relations’ that stretch over boundaries and are not territorially fixed have been accorded an increased attention in the regional studies. This research investigates the region building process in the Alpine region and analyses different agenda-settings pursued by various stakeholders in the Alpine region. The study employs qualitative methods to analyse processes of region building by applying Paasi’s institutionalisation theory. The findings show the regional dynamics in the case under investigation and claims that the Alpine region has being transformed from a closed, bounded, territorially fixed entity to a relational one, based on not territorially fixed heterogeneous relations. This transformation leads from a relatively ‘fixed’ Alpine region (as defined by Alpine Convention) to a more fluid, unbounded and ‘fuzzy’ space – the Alpine macro-region, which is being developing.

Current research on network use in the labor market focuses primarily on network morphology. In this dissertation, I use hierarchical regression to examine the influence of macro-level context on network use in job search.This study relies on a unique data set that combines individual job history data for years 1985 through 2001 collected by the Survey of Social Dynamics and Migration in Russia (SSMDR) in 40 regions in 2001-2002 and corresponding regional macro-economic data published by Goskomstat, the State Statistical Committee of the Russian Federation.The first question of this study focuses on what accounts for the temporal and regional variation in personal network use in the Russian labor market. I find that, for the post-Soviet period, increase in network use in job search can be attributed to the growth of the private sector: Russian employers are becoming "social capitalists" who take advantage of the resources personal connections can offer. I also find that the chances of finding a new job through personal ties are higher in the regions with larger small business sectors and in the regions with lower economic performance.Next I examine how regional economic performance and unemployment affect workers' chances of getting new jobs in the private sector and in smaller size organizations by means of personal networks. The analysis shows that social networks do lead to employment in the private sector and that this relationship is positively affected by regional economic performance, but not by unemployment rate. I also find here that social ties are likely to lead to new jobs in small organizations, but that this relationship does not vary by region.Finally, I investigate how regional economic performance and unemployment rates affect wages for jobs found through personal networks in the private sector and in small organizations. I find that while the private sector rewards network use, small organizations do not. The relationship between network use and wages does not vary by region. That is, regional economic performance does not have an effect on this relationship.

The thesis of this dissertation is that automating domain remodeling in AI planning using macro operators and making remodeling more flexible and applicable can improve the planning performance and can enrich planning. In this dissertation, we present three novel ideas: (1) we present an instance-specific domain remodeling framework, (2) we recast the planning domain remodeling with macros as a parameter optimization problem, and (3) we combine two domain remodeling approaches in the instance-specific remodeling context. In the instance-specific domain remodeling, we choose the best macro-augmented domain model for every incoming problem instance using a predictor that relies on previously solved problem instances to estimate the macros to be added the domain. Training the predictor is achieved off-line based on the observed relation between the instance features and the planner performance in the macro-augmented domain models. On-line, the predictor is used to find the best remodeling of the domain based on the problem instance features. Our empirical results over a number of standard benchmark planning domains demonstrate that our predictors can speed up the fixed-remodeling method that chooses the best set of macros by up to 2.5 times. The results also show that there is a large room for improving the performance using instance-specific over fixed remodeling approaches. The second idea is recasting the domain remodeling with macros as a parameter optimization. We show that this remodeling approach can outperform standard macro learning tools, and that it can significantly speed up the domain evaluation preprocessing required to train the predictors in instance-specific remodeling, while maintaining similar performance. The final idea applies macro addition and operator removal to the instance-specific domain remodeling. While maintaining an acceptable probability of solubility preservation, we build a predictor that adds macros and removes original operators based on the instance’s features. The results show that this new remodeling significantly outperforms the macro-only fixed remodeling, and that it is better than the fixed domain models in a number of domains.