Thesis (Ph. D.)--Ohio State University, 2007.

Thesis (M. Sc.)--University of Canterbury, 2009. / Typescript (photocopy). Includes bibliographical references (leaves 91-92). Also available via the World Wide Web.

Publicación a texto completo no autorizada por el autor / Determina las relaciones entre el análisis de correspondencia y el análisis log-lineal para obtener un método combinado de análisis de Tablas de Contingencia, en este enfoque el análisis de correspondencia es visto como una técnica que sirve para la exploración de residuos en modelos log- lineales. Las relaciones entre los modelos log-lineales y el análisis de correspondencias permiten desarrollar una metodología apropiada para el análisis de variables cualitativas mediante la combinación de ambas técnicas. En este enfoque, los modelos log-lineales se usan para evaluar la importancia de los efectos de interacción destacando los efectos que presenten diferencias significativas y luego se emplea el análisis de correspondencias para presentar las relaciones entre las categorías de las variables, de tal manera que evidencie la complementariedad entre estos dos modelos. En este estudio se resume esta metodología de análisis y se presenta una aplicación donde se estudia la relación entre el nivel de satisfacción del usuario de la consulta externa del servicio de pediatría del Hospital Nacional de Policía. / Tesis

El creciente avance de las telecomunicaciones durante estos últimos años, ha despertado el interés de muchos investigadores en conseguir más y mejores técnicas que permitan establecer comunicaciones inalámbricas robustas y que a su vez, soporten grandes flujos de información a alta velocidad. No obstante, la utilización de un canal inalámbrico para el envío y recepción de señales, introduce efectos indeseados tales como interferencia y ruido, debido a que entre el emisor y el receptor se interponen diferentes agentes que se comportan de aletoriamente. Es así como se han desarrollado métodos que han despertado gran interés entre los investigadores, debido a la obtención de resultados sorprendentes. Entre estos métodos destacan el algoritmo MAP y los códigos LDPC. El primero puede ser utilizado tanto en el aspecto de la ecualización de señales o en la decodificación de señales, mientras que el segundo se utiliza específicamente para codificar y decodificar señales. De este modo, el objetivo principal de esta memoria es unir estos dos algoritmos para que trabajen de forma iterativa, utilizando la información proporcionada por el ecualizador para mejorar el proceso de decodificación y a su vez, realizar el proceso de ecualización utilizando la información del decodificador. A este sistema retroalimentado se le llama Esquema Turbo. Además, se plantea el algoritmo para lenguajes no binarios. Para realizar este trabajo, primeramente se hizo un desglose de un sistema de comunicaciones actual, reconociendo cada uno de los bloques componentes, analizando su funcionamiento teórico, introduciendo modificaciones según corresponda y a partir de este análisis obtiene un diagrama de bloques que resume el funcionamiento general del algoritmo propuesto. Seguidamente se realizan pruebas a pequeña escala, utilizando un modelo de canal TDL Gaussiano de tres derivaciones, capaz de emular el comportamiento de una señal bajo interferencia y ruido, con el objetivo de demostrar el correcto funcionamiento del algoritmo y su convergencia, para posteriormente someter el esquema a transmisiones de grandes bloques de información, midiendo su desempeño bajo distintos escenarios y estudiando la tasa de símbolos errados (SER) variando la razón señal ruido (SNR). Los resultados obtenidos de esta investigación dicen que este algoritmo es capaz de lograr una corrección completa en señales pequeñas, incluso bajo condiciones en que de SNR se reduce hasta 5 dB para 7 iteraciones. Sin embargo, al momento de enviar bloques de información de tamaño mucho mayor (105 símbolos), el algoritmo presenta un piso de SNR de 10 dB, ya que al aumentarlo por sobre ese valor no se lograba mejoramiento de decodificación, debido a que se utilizó un esquema de codificación de bloques con matrices pequeñas. Esto se explica justamente porque la elaboración de matrices de codificación para códigos LDPC no binarios, para bloques grandes de información, es un tema de investigación que no se encuentra resuelto hasta la fecha.

The United States is the world’s largest poultry producer and exports about 18 percent of its total poultry production. It is also second largest exporter of broiler meats. Reports from the USDA predict that global import demand for poultry is expected to increase over the next 10 years, with the US accounting for 34% of the global poultry exports. The present study estimates the effects of exchange rate and US poultry export price on the quantity of poultry imports by the top five trading countries, namely Mexico, Canada, China, Hong Kong and Russia, during the period 1993 to 2012, using a double-log multiple regression model. Comparison of the effects across the countries was made possible with the incorporation of dummy variables for each country with Hong Kong serving as the baseline. The results demonstrated that the effect of exchange rate and poultry price, and per capita GDP on the quantity of poultry imported by Russia , Canada, and China is statistically different from Hong Kong and the rest of the countries in this study. Exchange rate appears to have a negative and statistically significant effect on US quantity of poultry exports. Export price and per capital GDP shows a positive and statistically significant impact on US poultry exports, although the result differs for individual countries. Overall, this study suggests that the effect of exchange rate and export price on the quantity of US poultry exports varies across countries. Key Words: Exchange rate, Double-log regression, US poultry price, Poultry exports

This thesis introduces two notions of motive associated to a log scheme. We introduce a category of log motives à la Voevodsky, and prove that the embedding of Voevodsky motives is an equivalence, in particular proving that any homotopy-invariant cohomology theory of schemes extends uniquely to log schemes. In the case of a log smooth degeneration, we give an explicit construction of the motivic Albanese of the degeneration, and show that the Hodge realization of this construction gives the Albanese of the limit Hodge structure.

Made available in DSpace on 2014-06-12T15:55:32Z (GMT). No. of bitstreams: 2 arquivo2239_1.pdf: 10396822 bytes, checksum: dccadd3953f0ad40fde3170750e97c82 (MD5) license.txt: 1748 bytes, checksum: 8a4605be74aa9ea9d79846c1fba20a33 (MD5) Previous issue date: 2010 / Fundação de Amparo à Ciência e Tecnologia do Estado de Pernambuco / A atual sociedade demanda métodos cada vez mais eficazes para proteger o acesso a instalações, a bens e a informações. Este controle, que pode ser entendido como um problema de identificação, é tradicionalmente realizado através do uso de nomes ou códigos de usuário, senhas, chaves e cartões. Entretanto, cartões e chaves podem ser perdidos, roubados ou copiados e nomes de usuário e senhas podem ser esquecidos, compartilhados ou até adivinhados. Métodos biométricos utilizam características físicas ou comportamentais possuídas pelos indivíduos para realizar a identificação. O uso de informações biométricas tem recebido grande atenção devido ao fato de que tais características não podem ser (ou dificilmente são) esquecidas, compartilhadas ou modificadas, sem assumir certo risco. Entre os diversos métodos biométricos, os sistemas de identificação baseados no reconhecimento da íris humana são frequentemente citados como uma das biometrias mais precisas. A presente dissertação descreve um sistema de reconhecimento de íris, baseado no modelo proposto por Libor Masek, composto pelas etapas de segmentação, normalização, extração de características (e codificação) e comparação. Modificações, em relação ao modelo original, foram propostas para as etapas de segmentação e extração de características. O uso de filtros log-Gabor 2D é investigado e os resultados alcançados são comparados com os obtidos pelo método sugerido por Masek. Um novo esquema para a etapa de segmentação também é apresentado. O método proposto combina técnicas de contorno ativo (AC) ao algoritmo Pulling-Pushing (PP desenvolvido por Zhaofeng He), dando origem ao modelo PP AC. Os resultados obtidos neste trabalho corroboram a idéia de que o reconhecimento de indivíduos através da íris possui ótima precisão, constituindo uma excelente escolha para a construção de sistemas de identificação

The report is part of the development project run by Ingenjörsfirman Geotech AB, where some of the field investigations were carried out in collaboration with Bohusgeo AB. The aim of the thesis is to do field investigations with the new Vane Geotech Nova (Nova-vane) equipment, and compare the results with known and proven methods. The vane test is an in situ method that determines the undrained shear strength, for example for clay soils. Eight tests have been performed with the Nova-vane and they have been compared with the Nilcon-vane, an electronic Nilcon-vane, a CPT-probing and Direct shear tests in the laboratory. Parameter determination has been done with CRS-tests and ordinary laboratory routine examination. According to earlier research, the time for failure, waiting time and rotational speed are parameters that have been tested in order to know how they affect the evaluated shear strength in clay soils. Previous investigations from 1950 to 1990 show that the parameters mentioned above have a major impact on the evaluated shear strength. When comparing already known evaluation methods for the undrained shear strength, it has been found that the Nova-vane generates shear strength measurements are far too low. The reason why the Nova-vane shows lower undrained shear strengths than other methods may be due to a number of factors. The analysis of the tests shows that the rotation speed has not been constant, as the rotation is becoming faster and faster, the time to failure varies between a few seconds to several minutes over the recommended time to failure. Another major factor is the deviating angle compared to the vertical plane, the vane has been driven down with. In the eight tests, the rotational speed has been a difficult parameter to control. The electric engine used, a modified Geotech EVT 2000 electric vane instrument, could only apply different energies and are normally used for 22 mm probe rods. Which specific applied force on the rods generating the correct rotational speed, depends entirely on the properties of the clay. When 36 mm probe rods are used together with the electric engine the rotational speed is even more difficult to control. According to the SGF recommended standard from 1993 states that the time to failure of a vane test should be between 2 and 4 minutes, from the time the vane is activated. It primarily depends on the speed of rotation, which means that the time for failure was also a difficult parameter to control. The angle with which the Nova-vane is driven increases with depth to about 30° at 30 meters. While the angle of CPT-probing increases only 5° at 30 meters. The driven angle could have significance to the shear strength, although it is difficult to conclude what kind. The Nova-vane is the only vane borer that can measure which angle it is driven down with. The increasing angle is assumed to depend on the geometry of the instrument and not on the properties of the clay. The conclusion is that the measurement with the Nova-vane gives lower values than the other methods. However, the range of values proves only small differences down to the depth of 20 meters. The system as a whole is working with an acceptable accuracy, although a stronger engine and a more precise control of the engine speed is needed. More tests have to be done to determine how the Nova-system can be designed and how the investigation should be performed to get equal values as other proven methods. / Rapporten är en del av ett utvecklingsprojekt som drivs av Ingenjörsfirman Geotech AB, där viss del av fältundersökningar utförts i samarbete med Bohusgeo AB. Examensarbetets roll har varit att utföra fältförsök med den nya vingförsöksutrustningen typ Vane Geotech Nova (Nova-vingsond) för att sedan sammanställa och jämföra resultaten med redan kända och väl beprövade metoder. Vingförsöksmetoden är en in situ metod som bestämmer den odränerade skjuvhållfastheten i kohesionsjord. Åtta försök med Nova-vingsond har utförts och jämförts med Nilcon-vingar och eldriven Nilcon-vinge samt CPT-sondering och direkta skjuvförsök i laboratorium. Även CRS-försök och rutinundersökning i laboratorium har gjorts, för parameterbestämning. Enligt undersökningar i tidigare forskning har: tiden till brott, väntetiden och rotationshastigheten varit parametrar som testats för att se hur de påverkar den utvärderade skjuvhållfastheten i lerjordar. Tidigare forskning utförd mellan 1950 och 1990, visar att nämnda parametrar har en stor inverkan på den utvärderade skjuvhållfastheten. Vid jämförelser mellan Nova-vingsonden och andra utförda försök inom provtagningslokalen ger Nova-vingsonden lägre odränerad skjuvhållfasthet, vilket kan bero på flertalet faktorer. Analysen av försöken påvisar att hastigheten inte har varit konstant utan att rotationen blir allt snabbare, tiden till brott varierar mellan några få sekunder till flera minuter över den rekommenderade tiden till brott. Samt att neddrivningsvinkeln är avvikande från det tänkta vertikalplanet. I de åtta försöken har rotationshastigheten varit en svårstyrd parameter. Den elektriska rotationsmotorn som har använts är en modifierad Geotech EVT 2000. Den elektriska är dimensionerad för sondstänger med en diameter på 22 mm och har en effektinställning. Vilken rotationshastighet som den specifika effekten genererar beror helt på lerans egenskaper, vilket gör att rotationshastighet är svår att ställa in på rotationsmotorn, speciellt med 36 mm sondstänger. Tiden till brott för vingförsöksmetoden ska enligt SGF standard från 1993 vara mellan 2 och 4 minuter från dess att vingen aktiverats. Den beror framför allt av rotationshastigheten, vilket medför att även tiden till brott var en svårstyrd parameter. Nova-vingsonden visar en växande neddrivningsvinkel mot djupet, vid 30 meter cirka 30°, i jämförelse med CTP-sonden som endast ökade med 5° på 30 meter.  Neddrivnings-vinkeln kan ha betydelse för den erhållna skjuvhållfastheten, dock är det en svår att bestämma hur. Nova-vingsonden är den enda vingförsöksmetod som har en lodenhet. CPT-sonden har också en lodenhet, den gav mindre utslag. Neddrivningsvinkeln bero förmodligen på instrumentets geometri och inte på lerans egenskaper. För att kunna styra rotationshastigheten och på så vis tiden till brott behövs en utvecklad utrustning. Främst krävs en starkare motor som rotationshastgeten går att ställa in på, övriga komponenter anses vara tillräckligt noggranna, dock behöver viss systemutveckling ske. Slutsatsen blir att de utförda mätningarna med Nova-vingsonden visar lägre resultat än övriga bestämningar. Då det är många parametrar som har inverkan på resultatet, behövs fler undersökningar göras för att fastställa hur Nova-systemet ska konstrueras och hur försök ska utföras för att erhålla likvärdiga resultat som övriga metoder.

Abstract Almost all the Finnish buildings were log constructed up to the beginning of the 20th century. Among these buildings one can see the finest expressions of Finnish wood architecture. Unfortunately, the tradition of the original and fascinating Finnish log architecture was broken in the beginning of the 20th century. At the same time a lot of knowledge and skills connected to log constructing were lost. In order to be able to continue the tradition of log architecture, more research work needs to be done. The rules that log constructions set to the architecture are still poorly known. Finnish State Railways was the most notable public builder in Finland in the end of the 1900s and constructed more than 4000 log buildings between years 1860–1950. Very high quality demands were set on the railway buildings and all the builders were highly experienced professionals. When studying traditional, professionally built log constructions, the railway stations are unique research subjects. In this thesis the log frames, structures and joints of 30 railway stations were documented. The main aim with the research was to find out how log structures affect the architectural form of the railway stations. As the results of the research it can be stated that in the architecture of the railway stations there are several factors which depend on log structures and which have had a powerful influence on the architectural form of railway stations and, thus, have created an identifiable building type for log stations. Due to the log and cross logging techniques, the most natural form of a log frame is a rectangular box, the height of which is the same everywhere and in which the openings obey certain rules. The rectangular box shape is so dominating in the railway stations that, on the basis of the log frame only, it is very difficult to place them to a certain architectural era or to categorize them according to a certain architectural style. The architectural originality has been created by different kinds of roof shapes, parting of the windows, sidings, decorations, porches and lean-to roofs. According to the data gathered from the railway stations, the architectural form of a log frame must be submitted to the rules of the log structure. Thus, the materialistic characteristics, the cross logging technique and the strength of the log structures are the factors that determine the ground plan, the masses and the openings of log buildings. Log as a construction material creates natural, small and simple buildings with undisturbed and balanced openings. The development work to be done on the basis of the rules of log structures can most naturally take place amongst handicraftsmen. The development of traditional log building still needs more research on historical building types, log joints and on the technical functioning of the log structures. / Tiivistelmä 1900-luvun alkuun saakka lähes kaikki maamme rakennukset olivat hirsirunkoisia. Näiden rakennusten joukosta löytyvät puuarkkitehtuurimme merkittävimmät saavutukset. Omintakeisen ja kiehtovan suomalaisen hirsiarkkitehtuurin perinne katkesi 1900-luvun alkupuolella. Samalla kadotettiin suuri joukko hirsirakentamiseen liittyvää tietoa ja taitoa. Perinnettä jatkavan hirsiarkkitehtuurin kehittymisen esteenä on tiedon puute. Emme tarkalleen tiedä, mitä sääntöjä perinteistä arkkitehtuuria ohjannut hirsirakenne asettaa arkkitehtoniselle muodonannolle. Valtion rautatiet oli maamme merkittävin julkinen rakentaja 1800-luvun loppupuolella ja rakensi vuosien 1860–1950 aikana yli 4 000 hirsirakennusta. Rakennuksille oli asetettu korkeat laatuvaatimukset, ja niiden rakentajina toimivat kokeneet ammattilaiset. Asemarakennukset ovat ainutlaatuinen kohde tutkittaessa perinteistä ammattimaista hirsirakentamista. Tutkimuksessa dokumentoitiin 30 asemarakennuksen hirsirunko, -rakenteet ja -liitokset. Dokumentointiaineiston pohjalta selvitettiin, miten hirsirakenteet ovat säädelleet asemarakennusten arkkitehtonista muodonantoa. Tutkimuksen tuloksena pystytään osoittamaan, että asemarakennuksissa on joukko hirsirakenteista johtuvia tekijöitä, jotka ovat merkittävästi vaikuttaneet rakennusten arkkitehtoniseen muodonantoon ja luoneet oman tunnistettavan hirsirunkoisten asemien rakennustyypin. Hirrestä ja lamasalvostekniikasta johtuvien tekijöiden vaikutuksesta hirsirungon luonnollinen muoto on tasakorkea suorakulmainen laatikko, jonka aukottamiseen on olemassa tiettyjä sääntöjä. Asemarakennuksissa on hirsirungon pelkistetty laatikkomuoto niin hallitseva, että rakennuksia on pelkän hirsirungon perusteella erittäin vaikeaa ajoittaa eri aikakausille tai luokittaa tiettyihin arkkitehtonisiin tyylikausiin. Rakennusten arkkitehtoninen yksilöllisyys on toteutettu vesikaton muodoilla, ikkunoiden jakopuitteilla, ulkoverhouksen struktuurilla, koristeaiheilla sekä keveillä kuisteilla ja katoksilla. Asemarakennuksista saadun tiedon mukaan hirsirungon muoto tulee alistaa hirsirakenteen ehtojen mukaiseksi. Täten rakennuksen pohjakaavaa, massoittelua ja aukotusta säätelevät hirren materiaaliominaisuudet, käytetty liitostekniikka ja hirsirakenteen lujuus. Hirrestä syntyy luonnollisesti pieniä, yksinkertaisia rakennuksia, joiden aukotus on rauhallista. Hirsirakenteen ehtojen pohjalta tapahtuva kehitystyö on luontevinta käsinveistotuotantotavan piirissä. Perinnettä jatkavan hirsiarkkitehtuurin kehittyminen vaatii vielä lisää tutkimusta historiallisista rakennustyypeistä, hirsiliitoksista ja hirsirakenteen teknisestä toiminnasta.

Abstract The objective of this work was to study the effect of adjustable process parameters on wood loss and bark removal in tumble drum debarking. The effect of capacity on the size distribution of the chips was studied in order to determine the optimal capacity for both the debarking and chipping sub-processes. The final aim was to propose a control strategy to optimise the processes by adapting their parameters according to the quality of the raw material. When the research started, earlier automation systems had focused on keeping the process alive, and economic values such as wood loss were not considered important. The process is usually controlled manually and shifts have different ways to manage it, which are based on trial and error. Bark removal in chemical pulp mills is usually higher than the values recommended in the literature, which in turn causes log breaking in the drum and increases wood loss. Even a small reduction in wood loss could have a substantial financial outcome. The lack of raw materials some time ago and the recession nowadays have highlighted the importance of more efficient log use. Data survey techniques were employed to reveal the interactions between drum variables from noisy measurements. Wood room data were analysed by modelling and deriving conclusions from the resulting parameters. In addition, log breaking and the size distribution of the chips were analysed under different process conditions. A pilot-scale drum was used to study residence time and the mechanical abrasion of logs. The results of this work indicated that the ratio of the volume of logs in the drum to capacity determines the residence time of the logs in the drum. Other variables influence the volume of logs in the drum, which together with capacity determine the residence time of logs in the drum, which affects wood loss and bark removal. The effect of capacity on the size distribution of the chips was not unambiguous, however, and it was therefore recommended to operate wood room at high capacity, because this reduces wood loss and increases annual production. The proposed control strategy adapts the residence time of logs in the drum to the quality of the debarked raw material by controlling the position of the closing gate. In addition, the control strategy adjusts the rotating speed of the drum using an open loop control. The results can be used to optimise the wood room process parameters. If the problem in wood room is excessive debarking, the residence time of the logs can be reduced and the rotating speed of the drum lowered. In this way the logs will be damaged less and wood loss will be reduced. Bark removal requirements in mechanical pulp mills are high, and the process parameters can be adapted to avoid problems in the subsequent processes due to excessive bark.