by Asano Masashi. / Thesis (M.B.A.)--Chinese University of Hong Kong, 1997. / Includes bibliographical references (leaves 97-98). / ABSTRACT --- p.ii / TABLE OFCONTENTS --- p.iii / LIST OF ILLUSTRATIONS --- p.iv / LIST OF TABLES --- p.v / ACKNOWLEDGEMENTS --- p.vi / Chapter / Chapter I. --- INTRODUCTION --- p.1 / Chapter II. --- METHODOLOGY --- p.7 / Chapter III. --- PROFIT MARGIN ANALYSIS --- p.22 / Chapter IV. --- SENSITIVITY ANALYSIS --- p.39 / Chapter V. --- COST ESTIMATE ANALYSIS --- p.48 / Chapter VI. --- INDUSTRY PROFITABILITY ANALYSIS --- p.76 / Chapter VII. --- CONCLUSION --- p.87 / APPENDIXES --- p.89 / BIBLIOGRAPHY --- p.97

Neste trabalho estendemos os modelos mistos semiparamétricos propostos por Zhang et al. (1998) para uma classe mais geral de modelos, a qual denominamos modelos mistos aditivos semiparamétricos com erros de contornos elípticos. Com essa nova abordagem, flexibilizamos a curtose da distribuição dos erros possibilitando a escolha de distribuições com caudas mais leves ou mais pesadas do que as caudas da distribuição normal padrão. Funções de verossimilhança penalizadas são aplicadas para a obtenção das estimativas de máxima verossimilhança com os respectivos erros padrão aproximados. Essas estimativas, sob erros de caudas pesadas, são robustas no sentido da distância de Mahalanobis contra observações aberrantes. Curvaturas de influência local são obtidas segundo alguns esquemas de perturbação e gráficos de diagnóstico são propostos. Exemplos ilustrativos são apresentados em que ajustes sob erros normais são comparados, através das metodologias de sensibilidade desenvolvidas no trabalho, com ajustes sob erros de contornos elípticos. / In this work we extend the models proposed by Zhang et al. (1998) to a more general class of models, know as semiparametric additive mixed models with elliptical errors in order to allow distributions with heavier or lighter tails than the normal ones. Penalized likelihood equations are applied to derive the maximum likelihood estimates which appear to be robust against outlying observations in the sense of the Mahalanobis distance. In order to study the sensitivity of the penalized estimates under some usual perturbation schemes in the model or data, the local influence curvatures are derived and some diagnostic graphics are proposed. Motivating examples preliminary analyzed under normal errors are reanalyzed under some appropriate elliptical errors. The local influence approach is used to compare the sensitivity of the model estimates.

CoordenaÃÃo de AperfeiÃoamento de Pessoal de NÃvel Superior / Conselho Nacional de Desenvolvimento CientÃfico e TecnolÃgico / Neste presente trabalho, faremos o estudo de importantes propriedades geomÃtricas e analÃticas de soluÃÃes de equaÃÃes diferenciais parciais elÃpticas totalmente nÃo-lineares do tipo: singulares e degeneradas. O estudo de processos de combustÃo que se degeneram ao longo do conjunto de anulamento da densidade de um gÃs, um caso particular de problemas do tipo "quenching", apresentam em sua modelagem equaÃÃes singulares que estÃo descritas neste trabalho. Nesta primeira parte iremos obter propriedades de uma soluÃÃo minimal, que vÃo desde o controle completo Ãtimo, atà a obtenÃÃo de estimativas de Hausdorff da fronteira livre singular. Por fim, iremos obter a regularidade Ãtima de soluÃÃes de equaÃÃes em que suas propriedades de difusÃo(elipticidade) se deterioram na ordem de uma potÃncia do seu gradiente ao longo do conjunto em que tal taxa de variaÃÃo se anula. / In this work we study important geometric and analytic properties to solutions of fully nonlinear elliptic partial differential equations, both singular and degenerate types. The study of combustion processes that degenerate along the null-set of the density of a gas, a particular case of quenching problems, present in their modeling, equations described in this work. In this first part we obtain properties of a minimal solution, since the complete optimal control until the Hausdorff estimates of the singular free boundary. Ultimately, we obtain the optimal regularity to equation solutions where their diffusion property (elipticity) deterorate in a power of their gradient along the set where such rate of variation nullifies.

Made available in DSpace on 2016-06-02T20:27:39Z (GMT). No. of bitstreams: 1 3703.pdf: 612745 bytes, checksum: 57763528b5a111b975b1122e35bbc887 (MD5) Previous issue date: 2011-06-16 / Universidade Federal de Minas Gerais / In this work, we extend some global L1 estimates proved by Bourgain-Brezis in the case of the de Rham complex on RN to the setup of local L1 estimates for elliptic complexes, namely, those associated to involutive elliptic structures spanned by a family of linearly independent smooth complex vector fields. In particular, we obtain a local version of Gagliardo-Nirenberg estimates for elliptic systems of vector fields. / Neste trabalho, estendemos algumas estimativas L1 provadas por Bourgain-Brezis no caso do complexo de de Rham em RN para o contexto local de estimativas L1 para complexos elíticos, a saber, aqueles associados a uma estrutura involutiva elítica gerada por uma família de campos vetoriais suaves e linearmente independentes. Em particular, obtemos uma versão local da desigualdade de Gagliardo-Nirenberg para um sistema de campos vetoriais elíticos.

The characterization of the local ill-posedness and the local degree of nonlinearity are of particular importance for the stable solution of nonlinear ill-posed problems. We present assertions concerning the interdependence between the ill-posedness of the nonlinear problem and its linearization. Moreover, we show that the concept of the degree of nonlinearity com bined with source conditions can be used to characterize the local ill-posedness and to derive a posteriori estimates for nonlinear ill-posed problems. A posteriori estimates are widely used in finite element and multigrid methods for the solution of nonlinear partial differential equations, but these techniques are in general not applicable to inverse an ill-posed problems. Additionally we show for the well-known Landweber method and the iteratively regularized Gauss-Newton method that they satisfy a posteriori estimates under source conditions; this can be used to prove convergence rates results.

This thesis introduces methods used in time-to-date analysis. It is written generally and so usable in dealing with any example. The thesis deals with problem of censoring, which means, that some observations occurred after the following, which is typical for the lifetime analysis. Methods mentioned in the thesis are nonparametric and parametric estimates of the survival function and their characteristics, and regression models, concretely Cox model and accelerated failure time model, which examine effect of the covariates on survival function. In the thesis is beside survival function presented hazard function, which express intensity of the analyzed event and cumulative hazard function, which is created as the name suggests by cumulative summation of the hazard function. Estimates of these functions are obtainable from survival function and for parametric estimate often exists formula resulting from parameters of used distribution. Empirical part of the thesis introduces influence of several different types and degrees of censoring on parametric and nonparametric estimates of the survival function, mean and median. The other empirical example is the usage of regression analysis on the data from the lungs cancer research made by Mayo Clinic.

The BEIR V preferred relative risk models and standard life-table techniques are used to project lifetime fatal cancer risk factors for average members of the Canadian population. Uncertainties associated with projections are evaluated for: (1) sampling variation (statistical error), (2) extrapolation of risks to low doses and low dose rates, (3) projection of excess lifetime cancer risks beyond the current periods of human observation in epidemiological studies, (4) the transfer of site-specific excess risk coefficients between populations with differing baseline cancer rates, and (5) the effect of differences in the age and sex distributions among occupations in the Canadian "radiation" workforce. Results are used to assess the applicability of the fatal cancer risk estimates recommended in ICRP publication 60 to the Canadian population. It was found that sampling variation, extrapolating to low doses and dose rates, projecting excess risks beyond current periods of observation, and the uncertainty in how to transfer site-specific excess risks between populations all cause substantial variations in lifetime cancer risk projections. Site-specific cancer risk projections may be expected to vary by factors of 2 to 5, depending on the source of uncertainty. Site-specific differences were found in the fatal cancer risk factors projected for "average" male and female workers among different occupations in the Canadian workforce. Site-specific worker averages differed by as much as a factor 3. Female average risk factors for digestive cancers were substantially higher than male workers, while male average risk factors tended to be higher for leukemia and respiratory cancer. Overall however, the majority of worker risk factors were within 2.5% of the site-specific projections for the workforce as a whole. The ICRP-60 nominal fatal cancer risk estimates, tissue weighting factors, and lifetime risk projections for prolonged radiation exposure were all in good agreement with equivalent values derived in this report for the Canadian population. In view of the uncertainties, the results suggest the ICRP estimated cancer risks are as good as any presently available and supports the use of the ICRP recommended values for the planning and regulation of radiation protection in Canada. / Thesis / Master of Science (MS)

In the increasingly competitive coal market, it is becoming more important for coal operators to develop mathematical models for surface mining which can estimate mining costs before the actual mining begins. The problem becomes even more acute with the new reclamation laws, as they affect surface coal mining methods, productivity, and costs. This study presents a computer simulator for a mountaintop removal type of surface mining operation. It will permit users to compare the costs associated with different overburden handling and reclamation plans. It may be used to minimize productivity losses, and, perhaps, to increase productivity and consequently to reduce operating costs through design and implementation of modified mountain top removal methods. / M.S.

Thesis (MBA)--Stellenbosch University, 2003. / ENGLISH ABSTRACT: This study is focussed on the all-important issue of estimation and forecasting in project management. Most project managers relate the kind of difficulties they have to go through in dealing with superiors that come up with a targeted deadline for a project with a given budget but with little or no regard to due process of arriving at both duration and cost. Many projects are doomed to overrun both schedule and cost from the word go because either the project manager is not equipped to give accurate estimates or is too timid to upset the boss by saying they do not believe the project would be completed within the said duration and the given cost without compromising quality. It is well said in literature that the three pillars of any project completion are time, cost and performance. Two of the three can be predetermined and the third will normally be dependent on those two. Where there is generous time and a large budget, chances are that a good quality well considered and executed project will result. The opposite is also true. This study investigates literature on the subject and also seeks to look at the trends in estimating in practice, by way of questionnaire and physical inspections. The questions that it seeks to answer are: a) Are project managers in most companies formally qualified? b) What is their disposition to using computerised/statistical methods in estimating? c) Does the level of education affect the disposition of project managers to using computerised/statistical methods of estimation? And finally, d) What methods are project managers using to estimate duration, cost and schedule? The results of this survey show that most projects managers do not have formal training in project management. At best they have attended short courses in project management. The survey also shows that even the formally educated project managers do not think much of strictly scientific (computerised/statistical) methods of project estimation. They rather feel that getting input from experienced functional people on the separate parts of the project tends to produce more accurate estimates. There is a leaning therefore towards using past experience on similar projects to generate norms and expert advice from the relevant functions on the expected costs and durations of accomplishing the parts of the project on which they have experience. All the project managers that responded to the survey said they had formal documentation for their projects and that they kept copies of the documentation in a central place that was easily accessible. Physical checks of this fact revealed that 29% of the interviewees did not have central or well-kept databases of previous projects' information. / AFRIKAANSE OPSOMMING: Die studie plaas klem op die belangrike kwessie van skatting en voorspelling in projekbestuur. Die meeste projekbestuurders ondervind die tipiese probleem waar senior persone projekteikendatums en -begrotings neerlê, sonder behoorlike oorweging van die faktore wat tyd en koste beïnvloed. Baie projekte is dus, van die staanspoor, bestem om beide koste en tyd te oorskry. Die projekbestuurder is dalk nie behoorlik toegerus om akkurate skattings te maak nie, of andersinds ontbreek die moed om die senior persoon daarop te wys dat tydskaal- en koste-teikens slegs teen 'n verlaging in kwaliteitstandaarde bereik kan word. Die literatuur gee wye erkenning aan wat die die drie pilare van projekvoltooing is: tyd, koste, en prestasie. Twee hiervan kan vooraf vasgepen word, terwyl die oorblywende een dan 'n funksie van die ander twee word. Indien daar ruim tyd en fondse beskikbaar is, is daar 'n beter kans om goeie kwaliteit en 'n gunstige projekuitkoms te verseker. Die teenoorgestelde is ook waar. Hierdie studie ondersoek die literatuur in die vakgebied, en gee bepaalde aandag aan die tendense in vooruitskattingspraktyke deur van 'n vraeboog en fisiese ondersoeke gebruik te maak. Die vrae wat deur hierdie navorsing gestel word, is die volgende: a) Is projekbestuurders, in die meeste organisasies, formeelopgelei? b) Wat is hulle houding teenoor die gebruik van rekenaargebaseerde statistiese metodes vir vooruitskatting? c) Beïnvloed die vlak van opleiding hierdie houding teenoor die gebruik rekenaargebaseerde statistiese metodes? d) Watter metodes gebruik projekbestuurders om kostes, tyd, en skedules te bepaal? Die ondersoek bevind dat die meerderheid projekbestuurders nie formele opleiding in projekbestuur het nie. Ten beste het hulle slegs kort kursusse in projekbestuur bygewoon. Die studie bevind ook verder dat selfs formeel opgeleide projekbestuurders nie ten gunste van streng wetenskaplike (rekenaargebaseerde statistiese) skattingstegnieke is nie. Hulle voel dat die insette van ervare funksionele personeel wat op verskillende elemente van die projek werk, meer akkuraat is. Daar is dus 'n neiging om eerder vorige ondervinding op soortgelyke projekte te gebruik om norme vas te stel, en om kundige spesialiste se raad te gebruik by die skatting van koste en tydsduur met betrekking tot die gedeeltes van die projek waarin hulle spesialiseer. AI die projekbestuurders wat aan die vraeboog-opname deelgeneem het, het bevestig dat hulle formele dokumentasie rakende hulle projekte byhou, en dat kopieë op In toeganklike plek sentraal beskikbaar gehou word. Die fisiese ondersoek bevind dat 29% van die persone waarmee onderhoude gevoer is, nie projekrekords van vorige projekte op In sentrale databasis bewaar nie.

published_or_final_version / Management Studies / Master / Master of Business Administration