Coordinated pricing and inventory management

This dissertation mainly focuses on coordinated pricing and inventory management problems, where the related background is provided in Chapter 1. Several periodic-review models are then discussed in Chapters 2,3,4 and 5, respectively. Chapter 2 analyzes a deterministic single-product model, where a price adjustment cost incurs if the current selling price is changed from the previous period. We develop exact algorithms for the problem under different conditions and find out that computation complexity varies significantly associated with the cost structure. %Moreover, our numerical study indicates that dynamic pricing strategies may outperform static pricing strategies even when price adjustment cost accounts for a significant portion of the total profit. Chapter 3 develops a single-product model in which demand of a period depends not only on the current selling price but also on past prices through the so-called reference price. Strongly polynomial time algorithms are designed for the case without no fixed ordering cost, and a heuristic is proposed for the general case together with an error bound estimation. Moreover, our illustrates through numerical studies that incorporating reference price effect into coordinated pricing and inventory models can have a significant impact on firms' profits. Chapter 4 discusses the stochastic version of the model in Chapter 3 when customers are loss averse. It extends the associated results developed in literature and proves that the reference price dependent base-stock policy is proved to be optimal under a certain conditions. Instead of dealing with specific problems, Chapter 5 establishes the preservation of supermodularity in a class of optimization problems. This property and its extensions include several existing results in the literature as special cases, and provide powerful tools as we illustrate their applications to several operations problems: the stochastic two-product model with cross-price effects, the two-stage inventory control model, and the self-financing model.

Integrating Aeon and Interlibrary Loan

Presentation from the MARAC conference in Roanoke, VA on October 7–10, 2015. S16 - A Tectonic Shift: Using Aeon to Improve Operations in Archives and Special Collections

Responsabilidade social: visões e perspectivas de empresários do segmento industrial do Rio Grande do Norte

The theme Corporate Social Responsibility is relatively recent both in the academic field and in the business practice. Because of the lack of socialization of experiences and precision of concepts, there are gaps regarding the understanding of the subject and, also, how to conduct operations. This study just seeks to investigate such matter, focusing applications of social responsibility in business. It takes, as empirical field, winners of the PSQT - SESI Prize of Quality at Work in Rio Grande do Norte (2002-2007) in order to systematize the various approaches on the issue, aiming to reveal subjective visions and perspectives of the theme. It is characterized as a qualitative study, carried out by structured interview. The universe was composed by 15 companies. It was used analysis of content categorical as an axis for the interpretation of the information. Three approaches guided the analysis: Business Ethics, (normative); Business & Society (contractual); Social Issues Management (strategic). The findings are related in three ways: 1) reasons for the CSR practice; 2) the results obtained; 3) the means of CSR. It was found that the award participation occurs, mostly, linked to SESI invitations, so, as an articulated movement of industrial corporations in Brazil it occurs, also, because of the organizational commitment with the society and the possibility of internal and social growths and because of the importance attributed to the report as an instrument of consultancy. There are no indicators to check impacts of organizational interventions, in spite of the existence of planning for the actions. Social responsibility appears as a tool to reinforce the organizational image and to increase satisfaction of the employees. There is a tendency of large and medium firms to a contractual commitment while the small and the micros firms are on strategic or normative level. The analysis of the perspectives of social responsibility future revealed trends towards for strategic approach

Otimização dos parâmetros de maquinagem no processo de fresagem

Cada vez mais, os principais objetivos na indústria é a produção a baixo custo, com a máxima qualidade e com o tempo de fabrico o mais curto possível. Para atingir esta meta, a indústria recorre, frequentemente, às máquinas de comando numérico (CNC), uma vez que com esta tecnologia torna se capaz alcançar uma elevada precisão e um tempo de processamento mais baixo. As máquinas ferramentas CNC podem ser aplicadas em diferentes processos de maquinagem, tais como: torneamento, fresagem, furação, entre outros. De todos estes processos, o mais utilizado é a fresagem devido à sua versatilidade. Utiliza-se normalmente este processo para maquinar materiais metálicos como é o caso do aço e dos ferros fundidos. Neste trabalho, são analisados os efeitos da variação de quatro parâmetros no processo de fresagem (velocidade de corte, velocidade de avanço, penetração radial e penetração axial), individualmente e a interação entre alguns deles, na variação da rugosidade num aço endurecido (aço 12738). Para essa análise são utilizados dois métodos de otimização: o método de Taguchi e o método das superfícies. O primeiro método foi utilizado para diminuir o número de combinações possíveis e, consequentemente, o número de ensaios a realizar é denominado por método de Taguchi. O método das superfícies ou método das superfícies de resposta (RSM) foi utilizado com o intuito de comparar os resultados obtidos com o método de Taguchi, de acordo com alguns trabalhos referidos na bibliografia especializada, o RSM converge mais rapidamente para um valor ótimo. O método de Taguchi é muito conhecido no setor industrial onde é utilizado para o controlo de qualidade. Apresenta conceitos interessantes, tais como robustez e perda de qualidade, sendo bastante útil para identificar variações do sistema de produção, durante o processo industrial, quantificando a variação e permitindo eliminar os fatores indesejáveis. Com este método foi vi construída uma matriz ortogonal L16 e para cada parâmetro foram definidos dois níveis diferentes e realizados dezasseis ensaios. Após cada ensaio, faz-se a medição superficial da rugosidade da peça. Com base nos resultados obtidos das medições da rugosidade é feito um tratamento estatístico dos dados através da análise de variância (Anova) a fim de determinar a influência de cada um dos parâmetros na rugosidade superficial. Verificou-se que a rugosidade mínima medida foi de 1,05m. Neste estudo foi também determinada a contribuição de cada um dos parâmetros de maquinagem e a sua interação. A análise dos valores de “F-ratio” (Anova) revela que os fatores mais importantes são a profundidade de corte radial e da interação entre profundidade de corte radial e profundidade de corte axial para minimizar a rugosidade da superfície. Estes têm contribuições de cerca de 30% e 24%, respetivamente. Numa segunda etapa este mesmo estudo foi realizado pelo método das superfícies, a fim de comparar os resultados por estes dois métodos e verificar qual o melhor método de otimização para minimizar a rugosidade. A metodologia das superfícies de resposta é baseada num conjunto de técnicas matemáticas e estatísticas úteis para modelar e analisar problemas em que a resposta de interesse é influenciada por diversas variáveis e cujo objetivo é otimizar essa resposta. Para este método apenas foram realizados cinco ensaios, ao contrário de Taguchi, uma vez que apenas em cinco ensaios consegue-se valores de rugosidade mais baixos do que a média da rugosidade no método de Taguchi. O valor mais baixo por este método foi de 1,03μm. Assim, conclui-se que RSM é um método de otimização mais adequado do que Taguchi para os ensaios realizados. Foram obtidos melhores resultados num menor número de ensaios, o que implica menos desgaste da ferramenta, menor tempo de processamento e uma redução significativa do material utilizado.

Simulating customer experience and word-of-mouth in retail: a case study

Agents offer a new and exciting way of understanding the world of work. In this paper we describe the development of agent-based simulation models, designed to help to understand the relationship between people management practices and retail performance. We report on the current development of our simulation models which includes new features concerning the evolution of customers over time. To test the features we have conducted a series of experiments dealing with customer pool sizes, standard and noise reduction modes, and the spread of customers’ word of mouth. To validate and evaluate our model, we introduce new performance measure specific to retail operations. We show that by varying different parameters in our model we can simulate a range of customer experiences leading to significant differences in performance measures. Ultimately, we are interested in better understanding the impact of changes in staff behavior due to changes in store management practices. Our multi-disciplinary research team draws upon expertise from work psychologists and computer scientists. Despite the fact we are working within a relatively novel and complex domain, it is clear that intelligent agents offer potential for fostering sustainable organizational capabilities in the future.

Contamination of port zone sediments by metals from Large Marine Ecosystems of Brazil

Sediment contamination by metals poses risks to coastal ecosystems and is considered to be problematic to dredging operations. In Brazil, there are differences in sedimentology along the Large Marine Ecosystems in relation to the metal distributions. We aimed to assess the extent of Al, Fe, Hg, Cd, Cr, Cu, Ni, Pb and Zn contamination in sediments from port zones in northeast (Mucuripe and Pecem) and southeast (Santos) Brazil through geochemical analyses and sediment quality ratings. The metal concentrations found in these port zones were higher than those observed in the continental shelf or the background values in both regions. In the northeast, metals were associated with carbonate, while in Santos, they were associated with mud. Geochemical analyses showed enrichments in Hg, Cd, Cu, Ni and Zn, and a simple application of international sediment quality guidelines failed to predict their impacts, whereas the use of site-specific values that were derived by geochemical and ecotoxicological approaches seemed to be more appropriate in the management of the dredged sediments. (C) 2012 Elsevier Ltd. All rights reserved.

Determination of subclinical metabolic disorders in transition dairy cows

Das Gesundheitsmanagement von Milchkühen hat in den vergangenen Jahren auf den landwirtschaftlichen Betrieben an Bedeutung gewonnen. Neben Präventionsmaßnahmen zur Gesunderhaltung der Tiere ist die frühzeitige und systematische Erkennung von Erkrankungen hierbei der Hauptbestandteil. Es zeigt sich vermehrt, dass vor allem Transitkühe verstärkt an Stoffwechselerkrankungen in sowohl klinischer als auch subklinischer Form erkranken. Letztere stellen ein hohes Risiko dar, zum einen weil subklinische Erkrankungen oftmals nur schwer oder gar nicht erkannt werden und zum anderen, weil sie in vielen Fällen die Grundlage für meist schwerwiegendere Folgeerkrankungen sind. In der vorliegenden Studie wird das Thema der Früherkennung von subklinischen Ketosen und der subakuten Pansenazidose behandelt. Verschiedene Methoden wurden unter praktischen Versuchsbedingungen auf ihre Tauglichkeit zur Krankheitserkennung hin geprüft. In einer ersten Studie wurde auf einem konventionellen Milchviehbetrieb ein Ketose-Monitoring bei frischlaktierenden Kühen ab Tag 3 postpartum durchgeführt. Insgesamt 15 Tiere waren an einer subklinischen Ketose erkrankt, was eine Aufkommensrate von 26% in den untersuchten Tieren bedeutete. Die Blutproben von insgesamt 24 Tieren wurden auf ihren IL-6-Gehalt untersucht. Von den untersuchten Tieren waren 14 Tiere erkrankt, 10 Tiere bildeten die gesunde Kontrollgruppe. Interleukin-6 wurde bestimmt, da dem Zytokin IL-6 in anderen Studien in Bezug auf Ketosen eine Rolle zugesprochen wurde. Die erwartete Erhöhung von IL-6 bei erkrankten Tieren konnte nicht festgestellt werden; die erkrankten Kühe zeigten vielmehr die niedrigsten IL-6 Werte der Studiengruppe. Insgesamt waren die IL-6 Konzentrationen auf einem niedrigen Niveau mit 27.2 pg/m l± 10.2. Es zeigte sich, dass die IL-6 Bestimmung im Blut hinsichtlich der Erkennung von subklinischen Ketosen nur eingeschränkt nutzbar ist. Es konnte ausschließlich eine schwache negative Korrelation zwischen Beta- Hydroxybutyrat (BHBA, Goldstandard für den Nachweis einer Ketose) und IL-6 detektiert werden. Zusätzlich zu den Blutanalysen wurde ebenfalls die tägliche Wiederkauaktivität mit dem „DairyCheck“ System bestimmt, welches kontinuierlich die charakteristischen Kaumuskelkontraktionen aufzeichnet und somit die Dauer des Wiederkäuens bestimmt werden kann. Es wurde geprüft, ob sich ketotische Tiere von nicht ketotischen Tieren hinsichtlich der täglichen Wiederkäuzeit unterscheiden. Milchkühe mit einer Ketose kauten im Schnitt 475 min/d ± 56 wieder, nach Genesung 497 min/d ± 48. Sie befanden sich somit im Durchschnitt immer unterhalb der gesunden Kontrollgruppe, welche 521 min/d ± 76 wiederkaute. Eine Korrelation zwischen der Wiederkauzeit und dem BHBA- Gehalt im Blut war nur sehr schwach ausgeprägt, nicht zuletzt da die Tiere generell eine hohe Variabilität in der Wiederkauaktivität zeigten. Bei einer weiteren Studie, ebenfalls auf einem Praxisbetrieb durchgeführt, wurde auf die Erkennung der subakuten Pansensazidose (SARA) fokussiert. Hierbei kam ein drahtloses, kommerziell verfügbares Bolussystem zum Einsatz, welches den pH Wert kontinuierlich im Retikulorumen misst. Es macht die Erkennung einer SARA auch unter Praxisbedingungen ohne invasive Methoden wie der Punktion möglich. Das Bolussystem wurde 24 Milchkühen kurz vor der Abkalbung oral eingegeben, um den pH-Wert während der gesamten Transitphase messen und überwachen zu können. Während in der Trockenstehphase nur vereinzelte SARA Fälle auftraten, erlitt ein Großteil der untersuchten Tiere in der Frühlaktation eine SARA. Auf Grundlage von pH-Werten von laktierenden Milchkühen, wurde zusätzlich eine Sensitivitätsanalyse von verschieden, bereits eingesetzten Nachweismethoden durchgeführt, um die Tauglichkeit für die SARA-Diagnostik zu untersuchen. Es handelte sich hierbei zum einen um einen SARA-Nachweis unter Heranziehung von Einzelwerten, Fress- und Wiederkäuzeiten, sowie ausgewählten Milchinhaltsstoffen und der Milchmenge. Die Analyse ergab, dass nahezu alle Nachweismethoden im Vergleich zur Langzeitmessung nur eingeschränkt zur SARA-Diagnostik nutzbar sind. In einem weiteren Teil der Studie wurde eine Kotfraktionierung bei den gleichen Tieren durchgeführt, um damit SARA-Tiere auch mittels der Kotanalyse erkennen kann. Es konnte gezeigt werden, dass zum einen die Ration einen Einfluss auf die Kotzusammensetzung hat (Trockensteherration versus Ration für Laktierende) zum anderen aber auch, dass eine SARA die Zusammensetzung des Kotes verändert. Hierfür wurden Kotproben ausschließlich von laktierenden Kühen untersucht, sodass der Einfluss der Ration ausgeschlossen werden konnte. Erhöhte Faseranteile im Kot von SARA - Kühen gaben Hinweis auf eine verminderte Verdaulichkeit. Dabei erwies sich vor allem die Hemizellulose als guter Parameter, um auf eine SARA schließen zu können. Die Versuchsbedingungen ließen es ebenfalls zu, die pH-Verläufe der Tiere in der Frühlaktation zu untersuchen. Eine Clusteranalyse von pH-Werten der ersten 12 Tage postpartum zeigte, dass die untersuchten Tiere trotz gleicher Haltungs- und Fütterungsbedingungen unterschiedliche pH-Wert Verläufe entwickelten. So gab es eine Gruppe von Milchkühen, die den pH-Wert stabil halten konnte, während die restlichen pH-Abfälle in verschiedenen Verläufen und Intensitäten aufzeigten. Es konnte ebenfalls aufgezeigt werden, dass Tiere innerhalb der Testherde unterschiedliche Schweregrade der SARA entwickelten. Auch in dieser Studie wurde deutlich, dass Tiere scheinbar unterschiedliche Möglichkeiten haben, auf ihre Umwelt zu reagieren, bzw. suboptimalen Bedingungen entgegenwirken zu können. Zusammengefasst wurden verschiedene Methoden zur Ketose- und SARA- Erkennung geprüft, von denen nur einzelne für die Praxis zu empfehlen sind. Die Variabilität der Tiere, sowohl bei der Ausprägung der Erkrankungen als auch bei den gemessenen Parametern verdeutlicht die Notwendigkeit, diese im Herden- und Gesundheitsmanagement in Zukunft stärker zu berücksichtigen.

Spacecraft Collision Probability Calculation from Conjunction Data Message and High Precision Orbit Propagator

The increasing number of Resident Space Objects (RSOs) is a threat to spaceflight operations. Conjunction Data Messages (CDMs) are sent to satellite operators to warn for possible future collision and their probabilities. The research project described herein pushed forward an algorithm that is able to update the collision probability directly on-board starting from CDMs and the state vector of the hosting satellite which is constantly updated thanks to an onboard GNSS receiver. A large set of methods for computing the collision probability was analyzed in order to find the best ones for this application. The selected algorithm was then tested to assess and improve its performance. Finally, parts of the algorithm and external software were implemented on a Raspberry Pi 3B+ board to demonstrate the compatibility of this approach with computational resources similar to those typically available onboard modern spacecraft.

Modelling fatigue damage distribution to inform slip and cut policy for riser tensioner ropes

The paper describes a method whereby the distribution of fatigue damage along riser tensioner ropes is calculated, taking account of heave motion, set tension, system geometry, tidal range and rope specification. From these data the distribution of damage along the rope is calculated for a given time period using a Miner’s summation method. This information can then be used to help the operator decide on the length of rope to ‘slip and cut’ whereby a length from the end of the rope is removed and the rope moved through the system from a storage drum such that sections of rope that have already suffered significant fatigue damage are not moved to positions where there is another peak in the distribution. There are two main advantages to be gained by using the fatigue damage model. The first is that it shows the amount of fatigue damage accumulating at different points along the rope, enabling the most highly damaged section to be removed well before failure. The second is that it makes for greater efficiency, as damage can be spread more evenly along the rope over time, avoiding the need to scrap long sections of undamaged rope.

Mechanistic approach to predict real machining time for milling free-form geometries applying high feed rate

An accurate estimate of machining time is very important for predicting delivery time, manufacturing costs, and also to help production process planning. Most commercial CAM software systems estimate the machining time in milling operations simply by dividing the entire tool path length by the programmed feed rate. This time estimate differs drastically from the real process time because the feed rate is not always constant due to machine and computer numerical controlled (CNC) limitations. This study presents a practical mechanistic method for milling time estimation when machining free-form geometries. The method considers a variable called machine response time (MRT) which characterizes the real CNC machine`s capacity to move in high feed rates in free-form geometries. MRT is a global performance feature which can be obtained for any type of CNC machine configuration by carrying out a simple test. For validating the methodology, a workpiece was used to generate NC programs for five different types of CNC machines. A practical industrial case study was also carried out to validate the method. The results indicated that MRT, and consequently, the real machining time, depends on the CNC machine`s potential: furthermore, the greater MRT, the larger the difference between predicted milling time and real milling time. The proposed method achieved an error range from 0.3% to 12% of the real machining time, whereas the CAM estimation achieved from 211% to 1244% error. The MRT-based process is also suggested as an instrument for helping in machine tool benchmarking.

AnSBBR Applied to a Personal Care Industry Wastewater Treatment: Effects of Fill Time, Volume Treated Per Cycle, and Organic Load

A study was performed regarding the effect of the relation between fill time, volume treated per cycle, and influent concentration at different applied organic loadings on the stability and efficiency of an anaerobic sequencing batch reactor containing immobilized biomass on polyurethane foam with recirculation of the liquid phase (AnSBBR) applied to the treatment of wastewater from a personal care industry. Total cycle length of the reactor was 8 h (480 min). Fill times were 10 min in the batch operation, 4 h in the fed-batch operation, and a 10-min batch followed by a 4-h fed batch in the mixed operation. Settling time was not necessary since the biomass was immobilized and decant time was 10 min. Volume of liquid medium in the reactor was 2.5 L, whereas volume treated per cycle ranged from 0.88 to 2.5 L in accordance with fill time. Influent concentration varied from 300 to 1,425 mg COD/L, resulting in an applied volumetric organic load of 0.9 and 1.5 g COD/L.d. Recirculation flow rate was 20 L/h, and the reactor was maintained at 30 A degrees C. Values of organic matter removal efficiency of filtered effluent samples were below 71% in the batch operations and above 74% in the operations of fed batch followed by batch. Feeding wastewater during part of the operational cycle was beneficial to the system, as it resulted in indirect control over the conversion of substrate into intermediates that would negatively interfere with the biochemical reactions regarding the degradation of organic matter. As a result, the average substrate consumption increased, leading to higher organic removal efficiencies in the fed-batch operations.

Influence of egg pre-storage heating period and storage length on incubation results

This experiment aimed at evaluating the influence of different heating times of settable eggs of Cobb 500 (R) broiler breeders before submitting them to different storage periods on egg weight loss, embryo mortality, and hatchability. A total number of 1,980 eggs were distributed in a completely randomized experimental design with a 3 x 3 factorial arrangement, comprising nine treatments with 22 replicates of 10 eggs each. The following factors were analyzed: pre-storage heating periods (0, 6, 12 hours at 36.92 degrees C) and storage periods (4, 9, 14 days at 12.06 degrees C). After storage, eggs were incubated under usual conditions, and were transferred to the hatcher at 442 hours of incubation. Eggs were weighed before heating, incubation, and transference to determine weight loss. Partial hatchability was determined at 480 hours, and total hatchability at 498 hours of incubation. Embryo mortality was determined in non-hatched eggs. It was concluded that heating eggs for six hour before storage improves incubation results as it decreases incubation length and late embryo mortality, therefore its use can be indicated in commercial operations. Storing eggs for 14 days and pre-heating for 14 days and pre-heating for 12 hours severely impair incubation results, and therefore are not recommended.

Transformation of external sensilla to chordotonal sensilla in the cut mutant of Drosophila assessed by single-cell marking in the embryo and larva

The cut gene of Drosophila melanogaster is an identity selector gene that establishes the program of development and differentiation of external sense organs. Mutations in the cut gene cause a transformation of the external sense organs into chordotonal organs, originally assessed by the use of immunostaining methods [Bodmer et al. (1987): Cell, 51:293-307]. Because of evidence that axonal projections of the transformed neurons within the central nervous system are not completely switched in cut mutants, the transformation of the four cells making up a sense organ was reassessed using single-cell staining with fluorescent dye and differential interface contrast (DIC) microscopy of the embryo and larva. The results provide strong evidence that all cells of the sense organs are completely transformed, exhibiting the morphologies and organelles characteristic of chordotonal sense organs. A comparison of the structures of external sense organs and chordotonal organs indicates that a number of the differences could be due to the degree of development of common structures, and that cut or downstream genes modulate effector genes that are normally utilized in both receptor types. The possible derivation of insect chordotonal and external sense organs from a receptor type found in crustaceans is discussed in the light of arthropod phylogenetics and the molecular genetics of sense organ development. (C) 1997 Wiley-Liss, Inc.

IS SPLENECTOMY A DYSLIPIDEMIC INTERVENTION? EXPERIMENTAL RESPONSE OF SERUM LIPIDS TO DIFFERENT DIETS AND OPERATIONS

Spleen removal may be recommended during organ transplantation in ABO-incompatible recipients as well as for hypoperfusion of the grafted liver, besides conventional surgical indications, but elevation of serum lipids has been observed in certain contexts. Aiming to analyze the influence of two dietary regimens on lipid profile, an experimental study was conducted. Methods: Male Wistar rats (n = 86, 333.0 +/- 32.2 g) were divided in four groups: group 1: controls; group 2: sham operation; group 3: total splenectomy; group 4: subtotal splenectomy with upper pole preservation; subgroups A (cholesterol reducing chow) and B (cholesterol-rich mixture) were established, and diet was given during 90 days. Total cholesterol (Tchol), high-density lipoprotein (HDL), low-density lipoprotein (LDL), very-low-density lipoprotein (VLDL), and triglycerides were documented. Results: After total splenectomy, hyperlipidemia ensued with cholesterol-reducing chow. Tchol, LDL, VLDL, triglycerides, and HDL changed from 56.4 +/- 9.2, 24.6 +/- 4.7, 9.7 +/- 2.2, 48.6 +/- 11.1, and 22.4 +/- 4.3 mg/dL to 66.9 +/- 11.4, 29.9 +/- 5.9, 10.9 +/- 2.3, 54.3 +/- 11.4, and 26.1 +/- 5.1 mg/dL, respectively. Upper pole preservation inhibited abnormalities of Tchol, HDL, VLDL, and triglycerides, and LDL decreased (23.6 +/- 4.9 vs. 22.1 +/- 5.1, P = 0.002). Higher concentrations were triggered by splenectomy and cholesterol-enriched diet (Tchol 59.4 +/- 10.1 vs. 83.9 +/- 14.3 mg/dL, P = 0.000), and upper-pole preservation diminished without abolishing hyperlipidemia (Tchol 55.9 +/- 10.0 vs. 62.3 +/- 7.8, P = 0.002). Conclusions: After splenectomy, hyperlipidemia occurred with both diets. Preservation of the upper pole tended to correct dyslipidemia in modality A and to attenuate it in subgroup B. (c) 2008 Wiley-Liss, Inc. Microsurgery 29:154-160, 2009.

Two-stage transvaginal cervical length screening for preterm birth in twin pregnancies

Aim: To compare cervical length (CL) at 18-21 and 22-25 weeks` gestation in twin pregnancies in prediction of spontaneous preterm delivery and to examine cervical shortening. Methods: Retrospective cohort study of CL measured at 18-21 and 22-25 weeks` gestation in twin pregnancies. Results: Receiver operating characteristics (ROC) curve revealed area of 0.64 (95% CI 0.53-0.75) and 0.80 (95% CI 0.72-0.88) for measurements at 18-21 and 22-25 weeks, respectively (P <= 0.001). Sensitivities of 33.3% and 23% and negative predicting value (NPV) of 97.3% and 86.8% for delivery at <28 and <34 weeks gestation were reached for measurements at 18-21 weeks. Sensitivities of 71.4% and 38.2% and NPV of 99.1% and 91.4% for delivery at <28 and <34 weeks` gestation were reached for measurements at 22-25 weeks. Cervical length shortening analysis showed an area under ROC curve of 0.81 (95% CI 0.73-0.89) and best cut-off was at >= 2 mm/week. Sensitivities of 80% and 60.8% and NPV of 98.9% and 90.6% for delivery at <28 and <34 weeks gestation were reached. Conclusions: In twin gestations, assessment of CL at 22-25 weeks is better than assessment at 18-21 weeks to predict preterm delivery before 34 weeks. Cervical shortening at a rate of >= 2 mm/weeks between 18 and 25 weeks gestation was a good predictor of spontaneous preterm birth in this high-risk population.