No presente trabalho são aplicadas ferramentas estatísticas quantitativas clássicas no setor de telefonia móvel brasileiro a fim de comparar os seus resultados. Uma metodologia é proposta para a aplicação destas ferramentas de um modo prático em uma de uma operadora de telefonia celular brasileira. Após são aplicados os métodos de previsão às séries históricas de demanda de acessos da telefonia móvel brasileira, estratificadas de acordo com as tecnologias utilizadas (TDMA, CDMA e GSM), no período de agosto de 2002 a julho de 2007. De acordo com o desempenho, são obtidos os modelos de cada método estatístico proposto. Entre estes, se escolheu aquele que melhor descreveu cada série e previsões foram geradas. Os modelos ARIMA apresentaram o melhor desempenho dentre os métodos aplicados nas séries estudadas. / The activity of planning ahead their systems in an appropriate manner is fundamental to telecommunication sector agents in order to manage the resources allocation and to meet the quality requirements in the provision of mobile telephony services. As the networks and production systems deployment takes time, there is the need of a service demand forecast. Therefore, in this study, classical quantitative statistics tools are applied to the Brazilian mobile telephony sector to compare their results. A methodology for the application of such tools in a practical way within a business environment of this sector is proposed. Afterwards, forecasting methods are applied to the time series referred to Brazilian mobile telephony demand, stratified according to the technologies (TDMA, CDMA and GSM) in the period from August 2002 to July 2007. The models of each statistical method proposed, based on the performance results, are obtained and, among these methods, it is chosen a model that best described each time series. The ARIMA model had the best performance among the methods applied in the time series studied and forecasts were made.

Diante do elevado nível de competição presente no cenário atual, torna-se indispensável a adoção de medidas de gestão capazes de priorizar e dirigir esforços na busca pela excelência no desempenho operacional das empresas. Assim, o objetivo deste trabalho é propor uma abordagem baseada em custos de produção e previsão de demanda voltada ao planejamento e controle da capacidade produtiva. Inicialmente, realizou-se o mapeamento dos artigos publicados em dezenove periódicos no período de 2002 a 2013, visando identificar abordagens relacionadas ao tema custos de produção e o processo de previsão de demanda. Com a finalidade de otimizar o planejamento da capacidade produtiva disponível, uma modelagem matemática, que utiliza o algoritmo generalized reduced gradient (GRG) não linear, é proposta. Por fim, é apresentado um modelo para o controle do desempenho operacional do sistema produtivo, fundamentado na avaliação de custos e no planejamento da capacidade de produção. Para avaliar a eficácia do modelo proposto, este foi aplicado em uma empresa de manufatura de materiais plásticos para a construção civil, em um sistema com múltiplos produtos e múltiplas máquinas. / Given the high level of competition present in the current environment, it is essential to adopt management measures in order to prioritize and direct efforts in the search for excellence in the company operational performance. Therefore, the objective of this study is to propose an approach based on production costs and demand forecasting focused on production capacity planning and control. A mapping of the articles published in nineteen journals, during the period of 2002 to 2013, was conducted to identify the different approaches related to production costs and demand forecasting. A mathematical modeling is proposed with the objective of optimizing the capacity planning using the nonlinear algorithm generalized reduced gradient (GRG). This study presented a model for controlling operational performance of the production system based on the evaluation of production costs and capacity planning. To evaluate the efficacy the model was applied to a manufacturing company of plastic materials for the construction, in a system with multi-products and multi-machines.

M.Comm. / As part of supply chain management, the handling of market demand information forms one of the most important concepts in any supply chain. One of the specific goals of supply chain management is to manage and co-ordinate the flow of information from the original source to the final customer. If consumer demand forms the activating element in the supply chain, it becomes clear that the process of demand planning can play an active role in improving the effectiveness of a supply chain. The correct management of information can thus greatly influence the level of integration, the responsiveness, level of customer service and value added to the end product. This is however not a one-sided approach where demand planning can be used as the tool to facilitate supply chain synchronization. The opposite effect can also be found that certain efforts to synchronize the supply chain can greatly improve the demand planning process. The fast moving consumer goods (FMCG) industry relies heavily on forecasted demand figures due to the structure of this industry 5. Developing demand forecasts forms a great part of the demand planning process and the accuracy, timely flow, interpretation and final format of the information is of the utmost importance. A well controlled forecasting process can form a solid foundation to address supply chain problems, reduce the level of wastage, increase the product value to the customer and improve the level of supply chain agility. With this background, the aim of this study will be: To explore the subject of Demand Planning in the synchronization of a FMCG supply chain. It will aim to show how an effective demand planning process can positively influence the supply chain management process and form an active element in supply chain synchronization. To investigate certain supply chain strategies on demand planning to indicate the level of integration between these two processes. In order to do this, a theoretical study needs to be done on Demand Planning and into the elements thereof. Within this structure it will be possible to formulate a structure to evaluate the concept of Demand Planning.

Esta tese apresenta uma sistemática de seleção dos especialistas mais consistentes e dos fatores de ajuste mais relevantes com vistas ao aprimoramento da acurácia da previsão de demanda gerada por métodos quantitativos. Para tanto, são testados sete modelos quantitativos: Médias Móveis (MM-3, MM-6 e MM-9), Suavização Exponencial Simples e Dupla e o modelo de Holt-Winters multiplicativo e aditivo. O modelo utilizado na previsão quantitativa foi aquele que gerou a melhor aderência aos dados e acurácia preditiva com base nos indicadores R2 e Erro Percentual Médio Absoluto (MAPE), respectivamente, extraídos mediante a quebra da série histórica na proporção 80% (banco de treino) e 20% (banco de teste) para cada produto. Com base nesse critério, tanto o leite UHT quanto o queijo mussarela foram modelados através da Suavização Exponencial Dupla (SED). Na sequência, especialistas e fatores utilizados para ajuste qualitativo da demanda foram selecionados de forma a reter somente os especialistas mais consistentes e os fatores mais influentes para tal fim. O método reteve os 5 especialistas mais consistentes dos 15 inicialmente entrevistados. Dos 23 fatores iniciais, apenas os 13 mais representativos foram retidos. Através da previsão corrigida para o leite UHT, o MAPE foi reduzido de 14,29% para 6,44%. Já previsão ajustada do queijo mussarela possibilitou reduzir o MAPE de 15,25% para 8,72%. / This thesis presents a systematic selection of the most consistent experts and most relevant adjustment factors aimed at improving the accuracy of forecasting demand generated by quantitative methods. For this, seven quantitative models are tested: Moving Averages (MM-3, MM-6 and MM-9), Single and Double Exponential Smoothing and Holt-Winters multiplicative and additive model. The model used in quantitative forecasting was one that generated the best adherence to data and predictive accuracy based on the indicators R2 and Mean Absolute Percentage Error (MAPE), respectively, extracted by breaking the time series in the ratio 80 % (workout bench) and 20% (test bank) for each product . Based on this criterion , both UHT milk and mozzarella cheese were modeled by Double Exponential Smoothing (SED). Further, experts and qualitative factors used to adjust demand were selected so to retain only the most consistent experts and the most influential factors for this purpose. The method retained the 5 most consistent experts of the 15 interviewed initially. Of the 23 initial factors, only the 13 most significant were retained. Through prediction corrected for UHT milk the MAPE was reduced from 14.29 % to 6.44 %. It had forecast adjusted mozzarella cheese possible to reduce the MAPE of 15.25% to 8,72.

The airline industry is a highly competitive market. Particularly after the major liberalization in the industry, the competition carried a step forward for all the concerned companies. Airline industry is a very dynamic market also; the expenses of a single operation may change substantially in a very short time period, which could be due to the fickle regulations or the fluctuation of oil prices. At any time, for a certain destination, a new airline company can start its operations and become a competitor, which can result in a market share loss for current operators. In such an environment, airline companies strive to maximize their revenues in every single flight. Load factors of a flight, classification, and pricing of airline tickets in a flight are determinative on the revenue of a single operation. In order to maximize revenue for its operations, companies should have a broad range of information about airline market conditions and passenger profile and preferences for every destination that they operate. Turkish Airlines, which is the largest airline company of Republic of Turkey and increasing its global awareness day by day, also strives to maximize its revenues per flight. In order to do so, Turkish Airlines needs to have suitable decisions regarding the capacity issues (determination of type and number of aircrafts, and flight frequencies) which will be then used worldwide in the different branches of company. These reasons lead to the increased need for the company to adopt practical total demand analysis in its operating destinations. The company aims to specify the factors, which are effective on the change of total demand for every destination. Turkish Airlines also strives to make most accurate demand forecasting in order to manage the peaks and troughs of the company’s flights. This project was brought to agenda in order to help this company to address this issue and was initiated with the purpose of making a total demand analysis for specified destinations. Destinations are from Stockholm to Athens, Thessalonica, New Delhi, Mumbai, Istanbul, Bangkok, and Dubai. This study is a quantitative study. Chosen demand factors, which constitute the total airline traffic between destination cities, are analyzed with respect to change in total demand. Various demand-forecasting methods are also applied in order to determine the best forecasting method for each destination, which is compatible with the quantitative approach. Results of the study are directly related to the airline market conditions of aforementioned destinations. Results are presented in four subheadings for every destination; overview of airline market, competition level of airline market, factors regarding the destination and demand forecasting for the destination. This study provides important information to Turkish Airlines regarding total demand analysis of specified destinations. Company learns the common factors (price, total amount of luggage, etc.) and the destination specific factors (language of crew, Frequent Flyer Programs, etc.) that are effective on the change of demand. Company may use this information in order to increase its market share (amount of passengers carried in a certain destination by Turkish Airlines from total demand) for specified destinations. With the help of this study, the company can make an accurate demand forecasting for Turkish Airlines’ future flights, which can be used in planning activities like determining the type of aircraft and flight frequency for a destination, and pricing of flight tickets through different passenger segments.

No description available.

Background: long term volume forecasting is important for logistics service providers for planning their capacity and taking the strategic decisions. At present demand is estimated by using traditional methods of averaging techniques or with their own experiences which often contain some error. This study is focused on filling these gaps by using machine learning approaches. The sample data set is provided by the organization, which is the leading manufacturer of trucks, buses and construction equipment, the organization has customers from more than 190 markets and has production facilities in 18 countries. Objectives: This study is to investigate a suitable machine learning algorithm that can be used for forecasting demand of outbound distributed products and then evaluating the performance of the selected algorithms by experimenting to articulate the possibility of using long-term forecasting in transportation. Methods: primarily, a literature review was initiated to find a suitable machine learn- ing algorithm and then based on the results of the literature review an experiment is performed to evaluate the performance of the selected algorithms Results: Selected CNN, ANN and LSTM models are performing quite well But based on the type and amount of historical data that models were given to learn, models have a very slight difference in performance measures in terms of forecasting performance. Comparisons are made with different measures that are selected by the literature review Conclusions. This study examines the efficacy of using Convolutional Neural Networks (CNN) for performing demand forecasting of outbound distributed products at the country level. The methodology provided uses convolutions on historical loads. The output from the convolutional operation is supplied to fully connected layers together with other relevant data. The presented methodology was implemented on an organization data set of outbound distributed products per month. Results obtained from the CNN were compared to results obtained by Long Short Term Memories LSTM sequence-to-sequence (LSTM S2S) and Artificial Neural Networks (ANN) for the same dataset. Experimental results showed that the CNN outperformed LSTM while producing comparable results to the ANN. Further testing is needed to compare the performances of different deep learning architectures in outbound forecasting.

京都大学 / 0048 / 新制・課程博士 / 博士(工学) / 甲第19979号 / 工博第4223号 / 新制||工||1653(附属図書館) / 33075 / 京都大学大学院工学研究科都市社会工学専攻 / (主査)教授 藤井 聡, 准教授 SCHMOECKER Jan-Dirk, 准教授 宇野 伸宏 / 学位規則第4条第1項該当 / Doctor of Philosophy (Engineering) / Kyoto University / DFAM

Time series prediction plays a pivotal role in various areas, including for example finance, weather forecasting, and traffic analysis. In this study, time series of historical sales data from a packaging manufacturer is used to investigate the effects that clustering such data has on forecasting performance. An experiment is carried out in which the time series data is first clustered using two separate approaches: k-means and Self-Organizing Map (SOM). The clustering is feature-based, meaning that characteristics extracted from the time series are used to compute similarity, rather than the raw time series. Then, A set of Long Short-Term Memory models (LSTMs) are trained; one that is trained on the entire dataset (global model), separate models trained on each of the clusters (cluster-based models), and finally a number of models trained on individual time series that are proportionally sampled from the clusters (single models). By evaluating the LSTMs based on Mean Absolute Error (MAE) and Mean Squared Error (MSE), we assess their consistency and predictive potential. The results reveal a trade-off between the consistency and predictive performance of the models. The global LSTM model consistently exhibits more stable performance across all predictions, showcasing its ability to capture the overall patterns in the data. However, the cluster-based LSTM models demonstrate potential for improved predictive performance within specific clusters, albeit with higher variability. This suggests that certain clusters possess distinct characteristics that allow for better predictions within those subsets of the data. Finally, the single LSTM models trained on individual time series, showcase even wider spreads of scores. The analysis suggests that the availability of training data plays a crucial role in the robustness (i.e., the ability to consistently produce similar results) of the forecasting models, with the global model benefiting from a larger training set. The higher variability in performance seen for the models trained with smaller training sets indicates that certain time series may be easier or harder to predict. It seems that the noise that comes with a larger training set can be either beneficial or detrimental to the predictive performance of the forecasting model on any individual time series, depending on the characteristics of that particular sample. Further analysis is required to investigate the factors contributing to the varying performance within each cluster. Exploring feature scores associated with poorly performing clusters and identifying the key features that contribute to better performance in certain clusters could provide valuable insights. Understanding these factors might aid in developing tailored strategies for cluster-specific prediction tasks.

No description available.