Master of Science / Department of Industrial and Manufacturing Systems Engineering / Todd Easton / Integer programs (IPs) are mathematical models that can provide an optimal solution to a variety of different problems. They have the ability to maximize profitability and decrease wasteful spending, but IPs are NP-complete resulting in many IPs that cannot be solved in reasonable periods of time. Cutting planes or valid inequalities have been used to decrease the time required to solve IPs. These valid inequalities are commonly created using a procedure called lifting. Lifting is a technique that strengthens existing valid inequalities without cutting off feasible solutions. Lifting inequalities can result in facet defining inequalities, the theoretically strongest valid inequalities. Because of these properties, lifting procedures are used in software to reduce the time required to solve an IP. This thesis introduces a new algorithm for synchronized simultaneous approximate lifting for multiple knapsack problems. Synchronized Simultaneous Approximate Lifting (SSAL) requires O(|E1|SLP_|E1|+|E2|,m + |E1|2) effort, where |E1| and |E2| are the sizes of sets used in the algorithm and SLP is the time to solve a linear program. It approximately uplifts two sets simultaneously to creates multiple inequalities of a particular form. These new valid inequalities generated by SSAL can be facet defining. A small computational study shows that SSAL is quick to execute, requiring fractions of a second. Additionally, applying SSAL inequalities to large knapsack problem enabled commercial software to solve faster and also eliminate off the initial linear relaxation point.

O objetivo deste trabalho é analisar a relação entre práticas de governança corporativa e desempenho de 370 empresas brasileiras de capital aberto no período de 1997 a 2007. O banco de dados utilizado foi construído a partir de dados extraídos dos relatórios Informativos Anuais (IAN), fornecidos pela Comissão de Valores Mobiliários (CVM) e pela base de dados do Economática®. Para considerar a causalidade reversa entre as variáveis de governança e desempenho e tendo em vista a presença de endogeneidade, utilizou-se um sistema de equações simultâneas composto por quatros equações. Neste sistema as variáveis de governança corporativa, estrutura de capital, estrutura de propriedade e desempenho foram consideradas endógenas. Os parâmetros do modelo foram estimados com a utilização do método de mínimos quadrados de dois estágios (MQ2E) e de mínimos quadrados de três estágios (MQ3E). O teste de Hausman sugere que as estimativas obtidas por meio de mínimos quadrados de três estágios são mais apropriadas para as inferências estatísticas. Os principais resultados indicaram que a diferença entre o direito de voto e o direito de fluxo de caixa em posse do maior acionista último é negativamente relacionada com o desempenho operacional da empresa. Isto indica que empresas com maior potencial de expropriação dos acionistas minoritários pelos majoritários tendem a apresentar menor desempenho. Esse resultado confirma a hipótese de que uma maior qualidade de governança corporativa impacta positivamente no desempenho da firma. Com a utilização da metodologia empregada, também foi possível analisar a influência do desempenho da firma sobre aspectos de governança corporativa. Os resultados apontaram que, quanto maior o desempenho maior a diferença entre o direito de voto e direito de fluxo de caixa do maior acionista último. Assim, os resultados evidenciam que a adoção de práticas de governança é importante para explicar o desempenho da firma, e também indicam a existência de inter-relação entre as variáveis testadas no modelo. / The objective of this study is to analyze the relationship between corporate governance practices and performance of 370 Brazilian companies traded in the period of 1997 to 2007. The used data base was created from extracted data of Informative Annual reports (IAN), provided by Comissão de Valores Mobiliários (CVM) and by the data base of Economática®. To consider the reverse causality between governance variables and performance and in view of the presence of endogeneity, we used a system of simultaneous equations consisting of four equations. In this system, the corporate governance variables, capital structure, ownership structure and performance were considered endogenous. The model parameters were estimated using the method of two-stage least squares (2SLS) and three-stage least squares (3SLS). The Hausman test suggests that the obtained estimates through the tree-stage least squares are more appropriate for statistical inferences. The main results indicate that the difference between voting right and cash flow right at the possession of the last largest shareholder is negatively related to the company\'s operating performance. This result indicates that firms with greater potential for expropriation of minority shareholders by majority shareholders tend to have lower performance which confirms the hypothesis that a higher quality of corporate governance has a positive impact on firm performance. Using this methodology, it was also possible to analyze the influence of the firm performance on aspects of corporate governance. The results support that, the higher the performance the greater the difference between voting rights and cash flow right to the last major shareholder. Thus, the results indicate that the adoption of governance practices is important to explain the performance and also indicate the existence of inter-relationship between the studied variables.

No description available.

Keywords. Simultaneous Interpreting; Compression; Strategy; Chinese (Cantonese)-English / Studying compression has both theoretical and practical implications for quality assessment in interpreting, interpreter training and interpreting practice. It could also provide us with concrete and palpable signs of the interpreting process. The findings of the study may also be applied to other language pairs, to the translation of subtitles, and to sign language interpreting. The research provides us with a new perspective to the objective assessment of some phenomena in SI, such as omissions, substitutions, deletions and simplifications. / The aim of this thesis is to investigate the use of compression strategy in simultaneous interpreting (SI) with reference to the language combination of Chinese (Cantonese) and English. The study uses concepts in linguistics, such as the Economy Principle, pragmatics, such as the Cooperative Principle and translation theories, such as Chernov's (1987/2004) categories of compression, Levy's (1967) MiniMax Principle, Descriptive Translation Studies (DTS), Skopos Theory, and "the Theory of Sense." The thesis addresses the following questions: (1) To what extent professional simultaneous interpreters resort to compression as a strategy in Chinese and English SI? (2) Is compression intrinsic to the process of SI and universal for all language combinations? If so, are Chernov's categories of compression applicable to Chinese and English SI? (3) What are the causes and factors of compression in SI? How factors such as contexts and situations, textual inferences and syntactic differences in the two language systems affect interpreters' use of compression as a strategy in SI? / The study concludes that interpreters use compression as a linguistic coping strategy to match with the speakers' delivery speed in the specific discourse environments and contexts of SI under temporal constraints. Compression is also a "labour-saving device" to cope with the cognitive load in the extreme conditions by simultaneous interpreters. It is also a stylistic device to achieve brevity and clarity for the target text, and an effective communicative device to realize the communicative act which benefits the addressees. Compression is found to be intrinsic to the process of SI which could be argued as a universal phenomenon and a universal strategy for all language combinations. / This research uses the quantitative and qualitative methods. It adopts a corpus-based approach, embracing discourse analysis, DTS, participant observation and simple statistics. This is achieved through compiling a parallel corpus of the performance of professional simultaneous interpreters in three meetings of Hong Kong Legislative Council (LegCo) in which the renderings are made from Cantonese into English, i.e. from A Language to B Language. It is found that compression occurs in 60% of the whole discourse in all the three meetings and compressions are used mainly as a linguistic coping strategy in specific situations and temporal constraints due to linguistic redundancy in human languages and specific characteristics of the spoken language. All of Chernov's categories of compression are found in Chinese (Cantonese) and English SI, such as syllabic compression, lexical compression, syntactic compression, semantic compression and situational compression. Pragmatic compression is also found in the data. The main factors of compression are linguistic redundancy of languages, specific characteristics of the spoken language, speech contexts and situations, different syntactic structures of Chinese and English, and interpreters' manipulation with delivery speed under stringent temporal constraints in the process of SI. Professional skills and experiences of individual interpreters and their delivery speed have also affected their use of compression in SI. / Wang, Yongqiu. / Adviser: Chan Sin-wai. / Source: Dissertation Abstracts International, Volume: 70-06, Section: A, page: 2019. / Thesis (Ph.D.)--Chinese University of Hong Kong, 2008. / Includes bibliographical references (leaves 303-320). / Electronic reproduction. Hong Kong : Chinese University of Hong Kong, [2012] System requirements: Adobe Acrobat Reader. Available via World Wide Web. / Electronic reproduction. [Ann Arbor, MI] : ProQuest Information and Learning, [200-] System requirements: Adobe Acrobat Reader. Available via World Wide Web. / Abstracts in English and Chinese. / School code: 1307.

In space, visual based relative navigation systems suffer from the harsh illumination conditions of the target (e.g. eclipse conditions, solar glare, etc.). In current Rendezvous and Docking (RvD) missions, most of these issues are addressed by advanced mission planning techniques (e.g strict manoeuvre timings). However, such planning would not always be feasible for Active Debris Removal (ADR) missions which have more unknowns. Fortunately, thermal infrared technology can operate under any lighting conditions and therefore has the potential to be exploited in the ADR scenario. In this context, this study investigates the benefits and the challenges of infrared based relative navigation. The infrared environment of ADR is very much different to that of terrestrial applications. This study proposes a methodology of modelling this environment in a computationally cost effective way to create a simulation environment in which the navigation solution can be tested. Through an intelligent classification of possible target surface coatings, the study is generalised to simulate the thermal environment of space debris in different orbit profiles. Through modelling various scenarios, the study also discusses the possible challenges of the infrared technology. In laboratory conditions, providing the thermal-vacuum environment of ADR, these theoretical findings were replicated. By use of this novel space debris set-up, the study investigates the behaviour of infrared cues extracted by different techniques and identifies the issue of short-lifespan features in the ADR scenarios. Based on these findings, the study suggests two different relative navigation methods based on the degree of target cooperativeness: partially cooperative targets, and uncooperative targets. Both algorithms provide the navigation solution with respect to an online reconstruction of the target. The method for partially cooperative targets provides a solution for smooth trajectories by exploiting the subsequent image tracks of features extracted from the first frame. The second algorithm is for uncooperative targets and exploits the target motion (e.g. tumbling) by formulating the problem in terms of a static target and a moving map (i.e. target structure) within a filtering framework. The optical flow information is related to the target motion derivatives and the target structure. A novel technique that uses the quality of the infrared cues to improve the algorithm performance is introduced. The problem of short measurement duration due to target tumbling motion is addressed by an innovative smart initialisation procedure. Both navigation solutions were tested in a number of different scenarios by using computer simulations and a specific laboratory set-up with real infrared camera. It is shown that these methods can perform well as the infrared-based navigation solutions using monocular cameras where knowledge relating to the infrared appearance of the target is limited.

There are numerous applications of game theory in the analysis of supply chains where multiple actors interact with each other in order to reach their own objectives. In this paper we review the use of non-cooperative game theory in inventory management within the newsvendor framework describing a single period inventory control model with the focus on horizontal interactions among multiple independent newsvendors. We develop a framework for identifying these types of horizontal interactions including, for example, the models with the possibility of inventory sharing via transshipments, and situations with substitutable products sold by multiple newsvendors. Based on this framework, we discuss and relate the results of prior research and identify future research opportunities.

Since weed killers were developed in the early part of the decade beginning in 1940, herbicides have been increasing in use until they are now widely used to control weeds . The application of herbicides has thus become a regular agricultural and ecological practice. Some of the herbicides are applied directly to soils and accompanied by a light irrigation to wash them into the soil in order to increase their efficiency . Some of these chemicals accumulate in the soil and should be removed in order to avoid toxic effect on succeeding plants (Klingman, 1961) . The toxic effect of herbicidal residues in soil and water to human subjects , livestock , and wildlife has been repor•ted (Paynter et al., 1960; Carpenter et al., 1961). When these materials are leached out of the soil , they may pollute the water supply. The movement of water and herbicides in soils is thus closely connected to both their herbicidal and toxic effect and is becoming a serious economic problem. It is the purpose of this investigation to study the simultaneous movement of water and herbicides in unsaturated soils by using the thermodynamic theory of irreversible processes. The adsorption of herbicides, the energy of adsorption, and the break-through curve will also be studied.

Doctor of Philosophy (PhD) / This thesis addresses the problem of building statistical models for multi-sensor perception in unstructured outdoor environments. The perception problem is divided into three distinct tasks: recognition, representation and association. Recognition is cast as a statistical classification problem where inputs are images or a combination of images and ranging information. Given the complexity and variability of natural environments, this thesis investigates the use of Bayesian statistics and supervised dimensionality reduction to incorporate prior information and fuse sensory data. A compact probabilistic representation of natural objects is essential for many problems in field robotics. This thesis presents techniques for combining non-linear dimensionality reduction with parametric learning through Expectation Maximisation to build general representations of natural features. Once created these models need to be rapidly processed to account for incoming information. To this end, techniques for efficient probabilistic inference are proposed. The robustness of localisation and mapping algorithms is directly related to reliable data association. Conventional algorithms employ only geometric information which can become inconsistent for large trajectories. A new data association algorithm incorporating visual and geometric information is proposed to improve the reliability of this task. The method uses a compact probabilistic representation of objects to fuse visual and geometric information for the association decision. The main contributions of this thesis are: 1) a stochastic representation of objects through non-linear dimensionality reduction; 2) a landmark recognition system using a visual and ranging sensors; 3) a data association algorithm combining appearance and position properties; 4) a real-time algorithm for detection and segmentation of natural objects from few training images and 5) a real-time place recognition system combining dimensionality reduction and Bayesian learning. The theoretical contributions of this thesis are demonstrated with a series of experiments in unstructured environments. In particular, the combination of recognition, representation and association algorithms is applied to the Simultaneous Localisation and Mapping problem (SLAM) to close large loops in outdoor trajectories, proving the benefits of the proposed methodology.

Peptide sequencing by surface-induced dissociation (SID) on a MALDI-Ion Mobility-orthogonal-TOF mass spectrometer is demonstrated. The early version of the instrument used for proof-of-concept experiments achieves a mobility resolution of approximately 20 and TOF mass resolution better than 200. Peptide sequences of four peptides from a tryptic digest of cytochrome c (ca. 1 pmol deposited) were obtained. The advantage of IM-SID-o-TOFMS is that a single experiment can be used to simultaneously measure the molecular weights of the tryptic peptide fragments (peptide mass mapping) and partial sequence analysis, (real time tandem mass spectrometry.) Optimization of the MALDI-IM-SID-o-TOF mass spectrometer for peptide sequencing is discussed. SID spectra obtained by using stainless steel, Au grids, and fluorinated self-assembled monolayers (F-SAM) on Au are compared. Optimum collision energies differ for the various surfaces. The fragmentation patterns observed for a series of peptides and protein digests using the Nd:YAG laser (355 nm) for MALDI ion formation and an FSAM surface for ion activation is compared to the fragmentation patterns observed for CID and photodissociation. The fragmentation patterns observed in all cases are strikingly similar. Photodissociation produced a greater abundance of ions resulting from side-chain cleavages. As a general rule optimized SID spectra contain fewer immonium ions than either photodissociation or CID. Evaluation of an instrument incorporating a new hybrid drift cell is discussed. Spectra for a digest of hemoglobin is compared to that acquired with an ABI 4700 TOF-TOF. The performance of the instrument is also evaluated using a micro-crystal Nd:YAG laser (355 nm) for MALDI operated at 400 Hz. Experiments were performed to determine the sensitivity and overall performance of the instrument. The reproducibility of the MS/MS spectra for gramicidin S is shown to be 94% run-to-run. The best mobility resolution obtained for a neat deposition of the dye Crystal Violet was 60 t/∆t. Sensitivity was tested with the peptide fibrinopeptide A (m/z 1537, AA sequence ADSGEGDFLAEGGGVR). Data acquired for sixty seconds with approximately sixty femtomoles deposited. Abundant [M+H]+ ions where observed as well as [M+H]+-NH3 ions. The S/N for this short run was insufficient to identify any SID fragments

This study examines the significance of corporate governance mechanisms during the corporate governance reform using a sample of 117 non-financial Japanese firms listed on the Tokyo Stock Exchange over the period 1989 to 2001. Japans prolonged recession brought about numerous reforms in post-bubble Japan. Although it is plausible to infer that the corporate governance system in Japan may have been a factor that led to the sustained recession in Japan, it is vague as to how deep and thorough the changes to Japanese corporate governance have been. The inference is that adverse impact of corporate governance may have been one of the factors that led to the sustained recession in Japan. Numerous proposals have been offered and some implemented in an attempt to fix problems exposed during the recession period in the 1990s. Remedies include instituting reforms to corporate governance by establishing new standards, punishing malpractice, and changing corporate board structures. Many Japanese firms look abroad for alternate governance mechanisms to integrate into their own system of control. As such, most reforms propose changes, for the most part, reflecting the American-style of corporate governance: alignment in incentives between top management and shareholders, board size reductions, and greater board independence to promote better monitoring and firmer discipline The significance of proposed changes to traditional Japanese corporate governance is examined in this study. Using a different econometric approach from that of previous studies, the relationship between board composition and firm performance is examined with a simultaneous framework of equations. The purpose of this empirical framework is to tackle potential endogeneity problems between board composition, governance and performance variables. Results show that: (1) there exists a significantly negative relationship between turnover of members of the board of directors and firm performance; (2) board size reduction is significant, but there is no evidence of consistent relationship between outside directorship and firm performance. (3) While keiretsu membership is generally relevant in linking board turnover and performance in Japan, board turnover is sensitive to performance in firms where ownership is concentrated than where ownership is dispersed. (4) President turnover, whether routine or non-routine, is unrelated to performance. Overall, results support that the entire board assumes responsibility for the firms performance and the 2SLS model is an effective estimator for estimating the relationship between board composition and firm performance.