Thesis (Ph. D.)--University of California, San Diego, 2000. / Vita. Includes bibliographical references.

Deux volets distincts de l'hydrologie sont abordés, prévision et prédétermination, au travers d'une problématique commune : le transfert à l'exutoire des bassins versants d' une information hydrologique distribuée. Dans le domaine de la prévision des crues, la technologie radar procure une information pluviométrique spatialement continue. Les hydrologues disposent ainsi en temps réel de la connaissance des champs de pluie, atout indéniable pour l'anticipation des crues notamment sur des petits bassins versants par le biais de la modélisation de la pluie en débit. Dans le cadre de la méthode AIGA d'alerte crues, développée au Cemagref, une modélisation mise en oeuvre à l'échelle du pixel de pluie fournit une cartographie des contributions de débit des pixels. Dans le domaine de la prédétermination, le Cemagref a développé la méthode SHYREG qui associe un modèle régionalisé de simulation de pluies horaires à une modélisation de la pluie en débit. Une estimation statistique régionale des pluies et des débits spécifiques de différentes durées, dans une large plage de fréquence (du courant à l'exceptionnel) peut ainsi être proposée et cartographiée. L'objectif du travail présenté est d'étudier et d'élaborer des méthodologies simples de transfert de ces deux informations débitmétriques discrétisées information temps réel pour le volet prévision et information statistique pour le volet prédétermination - à l'information débit à l'exutoire du bassin versant. La méthodologie met en oeuvre des informations spatiales et une modélisation de la pluie en débit. Pour répondre à l'objectif fixé, trois axes de travail sont développés. Le premier est l'étude du comportement d'un modèle pluie-débit simple développé pour être mis en oeuvre à la maille du km². On examine en particulier s'il satisfait les caractéristiques d'invariance et de parcimonie souhaitée pour une utilisation à la fois en reconstitution de crues et en simulation. Le second axe de travail concerne l'agglomération, en prédétermination, de l'information débit statistique connue au km² pour l'estimation des quantiles de débit à l' exutoire de bassins versants de superficie plus importante dans le cadre de la méthode SHYREG. Il s'agit de tenir compte de deux phénomènes hydrologiques distincts : l'abattement spatial de la pluie et le transfert dans le réseau hydraulique. Le troisième axe de travail concerne l'agglomération de l'information hydrologique distribuée pour la reconstitution des crues dans le cadre de l'outil AIGA d'alerte crues. Différentes modélisations sont proposées pour transférer à l'exutoire les contributions des débits modélisées aux pixels. / We address the routing of distributed hydrological information to the outlet of watersheds, in the fields of flood forecasting and flood prediction on ungauged watersheds in the French Mediterranean area.Flood forecasting can benefit of areal rainfall data provided in real-time by radar networks. This data used as an input to rainfall runoff models gives access to flood anticipation on small ungauged watersheds. Within the framework of the AIGA method, developed by CEMAGREF to provide floods alert, a rainfall-runoff model is implemented at the spatial resolution of the radar data, thus providing a map of the 1 km² pixel contributions to the runoff at the catchment outlet.Flood prediction consists of assessing the frequency of occurrence of floods of different given magnitude without reference to the times at which they would occur. The SHYREG flood prediction method, developed by Cemagref associates a regionalized rainfall model with a rainfall-runoff model. It provides grids of statistical estimates of rain and runoff for various duration and return periods. Our purpose is to study and work out simple methodologies to aggregate these two gridded hydrological data - real time information for the AIGA forecasting method and statistical data for the SHYREG prediction method to the catchments outlets. Our methodology implements distributed information and a rainfall-runoff model. We have first studied the behaviour of a simple rainfall-runoff model developed to be implemented in a gridded resolution (1 km² cells) for prediction as well as for forecasting purposes. We have checked that the model parameters show no redundancy and no link with the characteristics of the rainfall events. We have then addressed the question of the aggregation of gridded hydrological data. Within the SHYREG method, it consists of assessing statistical flow estimates at catchments outlets, knowing simulated flow distributions in each cell of the catchments. This aggregation would combine two distinct hydrological phenomena: areal reduction of rainfall and discharge attenuation in the channel network. Within the AIGA method, we have focused on the routing function of the rainfall-runoff model at the 1 km² cell scale, this scale being the first step of the runoff routing from the production area to the outlet of the catchment. We have then produced streamflow hindcasts for selected observed events using different routing function, within our rainfall-runoff model.

Magistro baigiamajame darbe pristatomas naujas kapitalo paskirstymo akcijų rinkoje modelis. Modelio teoriniam pagrindui naudojama modernioji portfelio teorija ir statistiniai prognozavimo metodai. Modelio praktinis panaudojimas pavaizduojamas Europos rinkų pavyzdžiu – peržiūrimi Nasdaq OMX internetiniame portale esantys 647 Europos akcijų pelningumo duomenys. Iš šių akcijų išrenkamos 7-ios, kurių pelningumas per paskutinius septynis metus buvo didžiausias. Atrinktos pelningiausių Europos firmų akcijos yra magistro baigiamojo darbo objektas. Darbo tikslas – išanalizavus ARIMA ir moderniosios portfelio teorijos kapitalo paskirstymo modelius suformuoti optimalų investicijų portfelį iš pelningiausių Europos akcijų, bei patikrinti hipotezę, kad taikant matematinius prognozavimo ir kapitalo paskirstymo modelius galima suformuoti portfelį, kurio pelningumo premijos ir rizikos santykis bus geresnis, nei palyginamųjų rinkų indeksų. Atsižvelgiant į darbo tikslą ir suformuotus uždavinius, pirmiausiai išrenkamos pelningiausios Europos akcijos ir trumpai apibūdinama išrinktų firmų veikla. Pristatomas kuriamo modelio teorinis pagrindas: supažindinama su ARIMA modelio koncepcija ir metodika, paaiškinama kapitalo paskirstymo atsižvelgiant į pelningumo ir rizikos santykį logika. Teorinis-loginis modelio formavimas pritaikomas praktiniame lygmenyje DNORD, MIC SDB, MOLS, ALFA, UIE, EKTA B, LEL akcijoms (santrumpų reikšmes galite rasti darbo pradžioje). Suformuoto investicijų portfelio... [toliau žr. visą tekstą] / In this concluding paper of master degree there is a new capital allocation in stock market model invented. The main principles of the modern portfolio theory and statistical forecasting are used as the theoretical basis of the model. The practical significance of the model is presented by using an example from the European markets – there are 647 European stock market shares analyzed. Seven shares that have the highest profitability over the last 7 years are selected as the components of the new portfolio. These shares are considered as the object of the research. The main goal of the research is to analyze ARIMA statistical forecasting and basic models of the modern portfolio theory and form an optimal investment portfolio consisting of the most profitable shares from the European market. The relevance of the research is verified by testing a hypothesis that mathematical models of forecasting and capital allocation can be applied to form a portfolio which would perform better than main indexes of the global markets in respect of the balance between profitability premium and risk. The process of the research is constructed as follows: selecting the most profitable shares from the European market, forming the theoretical background of the new model by introducing ARIMA forecasting and capital allocation principles, forming a new capital allocation and management model by combining ARIMA and H.Markowitz’s principles and practically using the new model for the selected (DNORD... [to full text]

abstract: In the U.S., high-speed passenger rail has recently become an active political topic, with multiple corridors currently being considered through federal and state level initiatives. One frequently cited benefit of high-speed rail proposals is that they offer a transition to a more sustainable transportation system with reduced greenhouse gas emissions and fossil energy consumption. This study investigates the feasibility of high-speed rail development as a long-term greenhouse gas emission mitigation strategy while considering major uncertainties in the technological and operational characteristics of intercity travel. First, I develop a general model for evaluating the emissions impact of intercity travel modes. This model incorporates aspects of life-cycle assessment and technological forecasting. The model is then used to compare future scenarios of energy and greenhouse gas emissions associated with the development of high-speed rail and other intercity travel technologies. Three specific rail corridors are evaluated and policy guidelines are developed regarding the emissions impacts of these investments. The results suggest prioritizing high-speed rail investments on short, dense corridors with fewer stops. Likewise, less emphasis should be placed on larger investments that require long construction times due to risks associated with payback of embedded emissions as competing technology improves. / Dissertation/Thesis / M.S. Sustainability 2011

Submitted by Diego Makasevicius Barbosa (diego.ufrj@yahoo.com.br) on 2017-11-24T18:24:17Z No. of bitstreams: 1 Trabalho Final (Assinada).pdf: 1388964 bytes, checksum: 33e74306625e467409ddc031936ccfeb (MD5) / Approved for entry into archive by Marcia Bacha (marcia.bacha@fgv.br) on 2017-12-08T16:45:39Z (GMT) No. of bitstreams: 1 Trabalho Final (Assinada).pdf: 1388964 bytes, checksum: 33e74306625e467409ddc031936ccfeb (MD5) / Made available in DSpace on 2017-12-08T16:46:06Z (GMT). No. of bitstreams: 1 Trabalho Final (Assinada).pdf: 1388964 bytes, checksum: 33e74306625e467409ddc031936ccfeb (MD5) Previous issue date: 2017-09-25 / This paper proposes to apply a similar framework adopted by Diebold and Li (2006) to forecast the Brazilian term structure of the US dollar-denominated interest rates, which have been done through the well-known three factors model developed by Nelson-Siegel. The methodology used to find the lambda factor, which drives the decay velocity of interest rates, was the rolling window optimization where for each forecast was calculated the lambda that minimizes the root mean square error (RMSE) of Nelson and Siegel fit. Furthermore, an autoregressive model was used to estimate the latent factors and, consequently, the interest rate. The results obtained were analogous to those found by Diebold and Li, where the authors verified a good predictive capacity for the model when compared to the random walk and other models used as benchmark. / O presente trabalho concentra-se em fazer um exercício de previsão da curva de cupom cambial futura similar ao proposto por Diebold e Li (2006) para as treasuries americanas, onde os autores utilizam um modelo econométrico de três fatores, no caso o conhecido Nelson e Siegel. A metodologia adotada para encontrar o fator λ (lambda), parâmetro este que rege a velocidade de decaimento da taxa do cupom cambial, foi uma otimização utilizando uma janela móvel, onde para cada instante t é observado qual o lambda que minimizaria a raiz do erro quadrático médio (REQM) do fit do modelo de Nelson-Siegel. Em seguida é conduzido um modelo autoregressivo para estimar os fatores latentes e consequentemente a taxa de cupom cambial para o exercício. O resultado obtido foi em linha com o encontrado por Diebold e Li onde os autores constataram uma boa capacidade preditiva para o modelo quando comparado ao passeio aleatório, nosso benchmark.

No description available.

Thesis (MScEng)--Stellenbosch University, 2014. / ENGLISH ABSTRACT: Tropical cyclones are synoptic scale rotating storms capable of generating intense wind speeds and rainfall with potentially devastating social and economic consequences. In addition to abnormally high winds and rainfall, the associated storm surge and extreme waves can lead to severe coastal erosion, damage to coastal property and inundation. A good understanding of the risk exposure to these events is therefore of great importance to planners and designers of coastal infrastructure in vulnerable regions. Probabilistic approaches have been routinely adopted for the calculation of extreme tropical cyclone induced wind speeds, with significant developments in these techniques over the last few decades. While the application of these approaches has become widely adopted in regions such as the North Atlantic, North Pacific and South Pacific Oceans, relatively little attention has been paid to the South-West Indian Ocean. This thesis focusses on the quantification of the risk exposure to tropical cyclones over the South-West Indian Ocean, using current state-of-the-art techniques. The primary results of the thesis are extreme wind speed maps at various return periods of interest for engineering design. Best track data for the South-West Indian Ocean, as archived by the Joint Typhoon Warning Centre (JTWC), has been used as the primary dataset forming the basis of this study. These data provide estimates of the location and intensity of historical tropical cyclones at six hourly intervals. Location data are provided as estimates of longitude and latitude of the eye, while intensity data are provided as estimates of the maximum sustained surface (10 m elevation) wind speed and/or minimum central pressure. The modelling of tropical cyclone wind fields has been carried out using both the Holland (1980) and the Willoughby et al. (2006) parametric wind field models. Using the limited information available in the best track data as input to the model, surface wind fields which reasonably resemble those of actual storms have been generated. Both considered parametric wind field models have been shown to yield reasonable wind speeds and directions when compared with measurements. Of the two considered models the Willoughby et al. (2006) model has been shown to provide the best fit to historical wind speed measurements. Extreme value analyses of tropical cyclone induced wind speeds based on historical data alone have been shown to lead to potentially large errors, owing to the small sample size of the historical data. This highlights the need to augment the historical database through a probabilistic approach. Largely following the methods described in Powel et al. (2005) and Emanuel et al. (2006), a synthetic track model for the South-West Indian Ocean has been developed. The objective of the synthetic track model is to simulate thousands of years of tropical cyclone tracks, thereby circumventing errors induced by small sample sizes in the available historical best track data. The synthetic track model developed as part of this study is a Markov chain model, capable of simulating track propagation and intensity evolution along the track, from track genesis through to termination. The model is purely statistical, based on properties derived from the historical best track data. Adjustments have however been made to account for physical limitations such as those imposed by the equator and the maximum potential intensity which an event can attain. The statistical characteristics of synthetic tracks have been shown to agree well with those of the historical population. Applying the Willoughby et al. (2006) wind field model along synthetic tracks has enabled the simulation of 5 000 years of tropical cyclone induced wind speeds at any location of interest in the South-West Indian Ocean. Applying calculations on a 1 degree geographical grid, wind speed maps corresponding to return periods of 50, 100, 200 and 500 years have been generated for the South-West Indian Ocean. Extreme wind speeds along coastal regions provide valuable input for the design of coastal infrastructure in the region. / AFRIKAANSE OPSOMMING: Tropiese siklone is sinoptiese orde roterende storms wat in staat is om aansienlike windspoed en reënval, tot gevolg te hê met potensiële vernietigende sosiale en ekonomiese gevolge. Benewens die abnormale sterk winde en hoë reënval kan die verwante stormdeinings en vloedgolwe lei tot ernstige kus-erosie, skade aan kusfront-eiendom en oorstromings. ‘n Goeie begrip van die risiko-blootstelling aan hierdie gebeurtenisse is daarom van groot belang vir die beplanners en ontwerpers van kus-infrastruktuur in kwesbare gebiede. As gevolg van die beduidende ontwikkeling van probabilistiese benadering tot die berekening van windspoed wat veroorsaak word deur ekstreme tropiese siklone, word hierdie tegnieke huidiglike op ‘n roetine basis aangewend. Terwyl die toepassing van hierdie benaderings wyd aanvaar word in gebiede soos die Noord-Atlantiese, Noordelike- en Suidelike Stille Oseaan, word relatief min aandag gegee aan die Suid-Westelike Indiese Oseaan. Hierdie tesis fokus op die kwantifisering van die risiko-blootstelling aan tropiese siklone in die Suid-Westelike Indiese Oseaan met die gebruik van die huidige gevorderdste tegnieke. Die primêre resultaat van die tesis is uiterste wind spoed kaarte vir ‘n verskeindenheid herhaal periodes wat van belang in vir engenieursontwerp. Beste roete-ata vir die Suid-Westelike Indiese Oseaan, soos voorsien deur die Gesametlike Tifoon Waarskuwing Sentrum (JTWC), is gebruik as die primêre data stel wat die basis vorm van hierdie studie. Hierdie data gee die beste skattings van die ligging (lengte- en breedtegraad), en intensiteit (maksimum volgehoue oppervlak (10m hoogte) wind spoed en/of sentrale druk tekort) van historiese tropiese siklone teen ses-uurlikse intervalle. Die modelering van tropiese sikloon windvelde was uitgevoer met die gebruik van die Holland (1980) en die Willoughby et al. (2006) parametriese windveldmodelle. Met die gebruik van beperkte inligting wat beskikbaar is in die beste roete data as invoer vir die model, was oppervlak wind velde gegenereer wat ‘n billike ooreenstemming het met die van werklike storms. Beide tegnieke se parametriese windveldmodelle is al bewys om redelike akkurate windspoed en windrigtings te lewer in vergelyking met waargenome waardes. Van die twee modelle het die Willoughby et al. (2006) model se resultate die beste ooreenstemming gewys met historiese wind spoed metings. Dit is al uitgewys dat uiterste waarde-analises van tropiese sikloon veroorsaakte windspoed moontlik kan lei tot groot foute in die resultate as gevolg van die klein monster-grootte van die historiese data. Dit beklemtoon die noodsaaklikheid om die historiese databasis aan te vul met behulp van probabilistiese metodes. Die metodes soos beskryf deur Powel et al. (2005) en Emanuel et al. (2006) is hoofsaaklik gebruik om ‘n sintetiese roete-model vir die Suid-Westelike Oseaan te ontwikkel. Die doelwit van die sintetiese roete model is om duisende jare se tropiese sikloonroetes te produseer, en in effek foute te vermy as gevolg van die gebruik van klein monster groottes van die beskikbare historiese beste roete data. Die sintetiese roete model wat tydens hierdie studie ontwikkel is, is ‘n Markov kettingmodel wat in staat is om die roete verspreiding asook die evolusie van intensiteit saam die roete te simuleer vanaf die onstaan tot die beëindiging van die sikloon se roete. Die model is suiwer statisties en is gebasseer op die eienskappe soos afgelei vanaf die historiese beste roete data. Aanpassings is gemaak om rekening te hou van die fisiese beperkings soos die wat opgelê word deur die ewenaar en die maksimum potensiële intensiteit wat ‘n sikloon kan bereik. Dit is voorgelê dat die statistiese einskappe van die sintetiese roetes goed saamstem met die van die historiese populasie. Die toepassing van die Willoughby et al. (2006) wind veld model langs die sintetiese roetes het dit moontlik gemaak om 5000 jaar se windspoed, wat veroorsaak is deur tropiese siklone, te genereer by enige ligging wat van belang is in die Suid-Westelike Indiese Oseaan. Met berekeninge wat op ‘n 1 grade geografiese ruitnet gedoen is, is windspoedkaarte vir herhaal periodes van 50, 100, 200 en 500 jaar opgestel vir die Suid-Westelike Indiese Oseaan. Die uiterste wind spoed in kusgebiede gee waardevolle invoer vir die ontwerp van kus-infrastruktuur in die omgewing.

En Guadeloupe, actuellement, 5,92% de la demande en énergie électrique sont couverts par la filière photovoltaïque et 3,14% par la filière éolienne soit 9,06% pour leur production cumulée selon le bilan 2015 de l’OREC (Observatoire Régional de l’Energie et du Climat). Selon le plan énergétique régional de prospection, la production cumulée du photovoltaïque et de l’éolien devrait représenter 14% du mix électrique en 2020 et 18% en 2030. Pour atteindre les 14% du mix électrique d’ici les cinq prochaines années, il va donc falloir entre autres, améliorer la prédictibilité pour un développement à un rythme soutenu de ces énergies. Ces travaux de recherches ont consisté à apporter de nouveaux résultats de performance de prévision de l’ensoleillement global à court terme et à donner une connaissance plus fine de la ressource sur trois stations en Guadeloupe. L’étude est basée sur une analyse et un modèle de prévision de l’ensoleillement, faisant intervenir des paramètres spatiaux et temporels. La littérature montre qu’un important nombre de sites est en général utilisé pour une analyse spatio-temporelle, ce qui impliquerait pour nous, de poser de multiples capteurs sur l’ensemble du territoire. Les coûts d’un tel système seraient considérables. Notre approche ici consistera à effectuer une analyse spatio-temporelle sur trois stations. Avec peu de stations et des distances non uniformes nous avons donc cherché à développer un modèle de prévision de l’ensoleillement à court terme en dépit de ces contraintes qui ne répondent pas à une approche classique. Le modèle est basé sur une méthodologie VAR (Vecteur Autorégressif) incluant des paramètres spatiaux et temporels. Une stratégie de sélection des variables est développée afin de sélectionner les prédicteurs (stations) utiles pour la prévision sur une localisation. Cette stratégie itérative permettra d’une part d’être plus proche de la réalité, d’autre part d’un point de vue algorithmique, la tendance des calculs sera plus rapide. En amont du développement du modèle, une étude de la variabilité spatio-temporelle de l’ensoleillement a permis de quantifier et caractériser de manière fine, les interactions dynamiques entre ces trois stations. Par comparaison avec les modèles de la littérature, notre modèle de prévision montre une bonne performance avec des valeurs de RMSE relative allant de 17,48% à 23,79% pour des horizons de prévisions de 5 min à 1h. Les méthodologies développées pourraient à terme offrir une opportunité d’assurer des garanties au gestionnaire du réseau. Si d'avenir des solutions de prévision performantes se généralisaient, cette opportunité permettrait d’ouvrir le marché au-delà du seuil de 30% imposé actuellement. / Currently in Guadeloupe, there is 5,92 % of the electric power request covered by the photovoltaic sector and 3,14 % by the wind sector which represents 9,06 % for their accumulated production, according to the OREC report (Regional Monitoring center of Energy and Climate). According to the regional energy plan, the accumulated production of the photovoltaic and the wind energy should represent 14 % of the electric mix in 2020 and 18 % in 2030. To reach the 14 % of the electric mix within the next five years, we need, among other things, to improve forecast for a sustained development of these energies. These research works consisted in bringing new performance results of short-term forecast of the global solar radiation and in giving a finer knowledge of the resource onto three stations in Guadeloupe. The study is based on an analysis and a forecast model of global solar radiation, by including spatial and temporal parameters. The literature shows that an important number of sites is generally used for a spatio-temporal analysis, which would imply for us, to put multiple sensors on the whole territory. The costs of such a system would be considerable. Our approach here will consist in making a spatiotemporal analysis on three stations. With few stations and not uniform distances, we, thus, tried to define a short-term forecast model of global solar radiation, in spite of these constraints which do not answer to a classic approach. The model is based on a methodology the VAR ( Autoregressive Vector) including spatial and temporal parameters. A strategy of selection of variables is developed to select useful predictors (stations) for the forecast on localization. This iterative strategy, on one hand will allow being closer to the reality, on the other hand to the point of algorithmic view, the trend of the calculations will be faster. Preliminarily, a study of the spatiotemporal variability of global solar radiation, allowed to quantify and to characterize in a fine way, the dynamic interactions between these three stations. Compared with the models of the literature, our forecast model shows a good performance with relative RMSE values going from 17.48 % to 23.79 % for horizons from 5 min to 1 hour. The developed methodologies could eventually offer an opportunity to assure guarantees to the network manager. If in the future the successful solutions of forecast became widespread, this opportunity would allow the opening of the market beyond the 30 % threshold imposed at present.

Today, digitalization is a key factor for businesses to enhance growth and gain advantages and insight in their operations. Both in planning operations and understanding customers the digitalization processes today have key roles, and companies are spending more and more resources in this fields to gain critical insights and enhance growth. The fast-food industry is no exception where restaurants need to be highly flexible and agile in their work. With this, there exists an immense demand for knowledge and insights to help restaurants plan their daily operations and there is a great need for organizations to continuously adapt new technological solutions into their existing processes. Well implemented Machine Learning solutions in combination with feature engineering are likely to bring value into the existing processes. Sales forecasting, which is the main field of study in this thesis work, has a vital role in planning of fast food restaurant's operations, both for budgeting purposes, but also for staffing purposes. The word fast food describes itself. With this comes a commitment to provide high quality food and rapid service to the customers. Understaffing can risk violating either quality of the food or service while overstaffing leads to low overall productivity. Generating highly reliable sales forecasts are thus vital to maximize profits and minimize operational risk. SARIMA, XGBoost and Random Forest were evaluated on training data consisting of sales numbers, business hours and categorical variables describing date and month. These models worked as base learners where sales predictions from a specific dataset were used as training data for a Support Vector Regression model (SVR). A stacking approach to this type of project shows sufficient results with a significant gain in prediction accuracy for all investigated restaurants on a 6-week aggregated timeline compared to the existing solution. / Digitalisering har idag en nyckelroll för att skapa tillväxt och insikter för företag, dessa insikter ger fördelar både inom planering och i förståelsen om deras kunder. Det här är ett område som företag lägger mer och mer resurser på för att skapa större förståelse om sin verksamhet och på så sätt öka tillväxten. Snabbmatsindustrin är inget undantag då restauranger behöver en hög grad av flexibilitet i sina arbetssätt för att möta kundbehovet. Det här skapar en stor efterfrågan av kunskap och insikter för att hjälpa dem i planeringen av deras dagliga arbete och det finns ett stort behov från företagen att kontinuerligt implementera nya tekniska lösningar i befintliga processer. Med väl implementerade maskininlärningslösningar i kombination med att skapa mer informativa variabler från befintlig data kan aktörer skapa mervärde till redan existerande processer. Försäljningsprognostisering, som är huvudområdet för den här studien, har en viktig roll för verksamhetsplaneringen inom snabbmatsindustrin, både inom budgetering och bemanning. Namnet snabbmat beskriver sig själv, med det följer ett löfte gentemot kunden att tillhandahålla hög kvalitet på maten samt att kunna tillhandahålla snabb service. Underbemanning kan riskera att bryta någon av dessa löften, antingen i undermålig kvalitet på maten eller att inte kunna leverera snabb service. Överbemanning riskerar i stället att leda till ineffektivitet i användandet av resurser. Att generera högst tillförlitliga prognoser är därför avgörande för att kunna maximera vinsten och minimera operativ risk. SARIMA, XGBoost och Random Forest utvärderades på ett träningsset bestående av försäljningssiffror, timme på dygnet och kategoriska variabler som beskriver dag och månad. Dessa modeller fungerar som basmodeller vars prediktioner från ett specifikt testset används som träningsdata till en Stödvektorsreggresionsmodell (SVR). Att använda stapling av maskininlärningsmodeller till den här typen av problem visade tillfredställande resultat där det påvisades en signifikant förbättring i prediktionssäkerhet under en 6 veckors aggregerad period gentemot den redan existerande modellen.

[pt] O varejo é conhecido como um ambiente competitivo e suscetível às mudanças de mercado. Assim, as marcas estão sempre se atualizando e inovando, principalmente em seu ponto de venda físico. A pandemia do Covid-19 acarretou uma aceleração digital no varejo de moda por conta do período em que as lojas físicas ficaram fechadas, colocando em questão a existência da loja física. O presente trabalho teve como objetivo repensar o design do ponto de venda físico de moda, em especial, o visual merchandising, a partir da pandemia do Covid-19. Para isso, utilizou-se de pesquisas bibliográfica e documental sobre temas relacionados à inovação no varejo, Covid-19, visual merchandising e tecnologias no ponto de venda; e das pesquisas de campo com consumidoras brasileiras e com especialistas da área de visual merchandising. Com base nas informações levantadas foi possível verificar que a loja física de moda e as estratégias relacionadas ao visual merchandising em um cenário pós-pandêmico estarão ligadas a três pilares: phygital, sustentabilidade e experiência de marca. A partir deles, foi possível fazer projeções para o design de varejo de moda. Esses pilares irão aumentar a percepção de valor em seus produtos e permitir uma maior fidelização de seus clientes. Desse modo, as marcas de moda precisam repensar suas estratégias no ponto de venda, o que evidencia a importância e o desafio para o profissional de visual merchandising nesse novo cenário. Assim, a loja física perde sua relevância como função exclusivamente transacional, sendo considerada como um local de conexão entre o consumidor e a marca. / [en] Retail business is a competitive environment and highly influenced by market changes. Thus, brands are always updating and innovating, especially as to their point of sale. The Covid-19 pandemic entailed a digital acceleration in the fashion retail industry due to the period in which physical stores were closed, raising the issue for their need. The present study aims to rethink the design at fashion physical stores using the VM approach within the context of necessary adaptations imposed by the Covid-19 pandemic. For this, we conducted bibliographic and documentary based research on topics related to innovation in retail, Covid-19, visual merchandising and point-of-sale Technologies. In addition, we conducted field surveys with Brazilian consumers and fashion retail specialists focusing on visual merchandising. Based on the information collected it was possible to verify that the physical fashion store and VM strategies in a post-pandemic scenario are linked to three axes: phygital, sustainability and retail experience. These three axes made it possible to project scenarios for designing fashion retail stores. They will add value to their products and allow for greater customer loyalty. For this reason, fashion brands need to rethink their strategies for sales at the point of sale, highlighting the importance and the challenge imposed on the VM professionals in this new scenario. Thus, the physical store is no longer a mere place for doing business, but a place where consumers and brands connect.