by Chan Ka Man. / Thesis (M.Phil.)--Chinese University of Hong Kong, 1999. / Includes bibliographical references (leaves 64-65). / Abstracts in English and Chinese. / ACKNOWLEDGMENTS --- p.iii / LIST OF TABLES --- p.iv / CHAPTER / Chapter I --- INTRODUCTION --- p.1 / Chapter II --- THE LITERATURE REVIEW --- p.4 / Chapter III --- METHODOLOGY / Forecast Pooling Technique / Modified Technique / Chapter IV --- MODEL SPECIFICATIONS --- p.16 / The Monthly Models / The Quarterly Model / Data Description / Chapter V --- THE COMBINED FORECAST --- p.32 / Pooling Forecast Technique in Case of Hong Kong / The Forecasts Results / Chapter VI --- CONCLUSION --- p.38 / TABLES --- p.40 / APPENDIX --- p.53 / BIBLIOGRAPHY --- p.64

The continuing evolution of urban travel patterns and changing policy goals and priorities requires that transportation researchers and practitioners improve their abilities to plan and forecast the demand for travel. Walking and bicycling - the primary forms of active travel - are generating increasing interest for their potential to reduce automobile use, save governmental and consumer costs, and improve personal and social health outcomes. Yet, current transportation planning tools, namely regional travel demand forecasting models, poorly represent these active travel modes, if at all. More broadly, travel models do an incomplete job of representing the decision-making processes involved in travel choices, especially those factors influencing walking and bicycling. In addition to limitations of data and statistical analysis methods, the research upon which modeling tools are based has yet to settle on a comprehensive theory of travel behavior that accounts for complex relationships around a variety of personal, social, and environmental factors. While modeling tools have explained travel primarily through economic theories, contributions from the geography and psychology fields prove promising. A few scholars have attempted to link these travel behavior explanations together, some with a focus on walking and bicycling, but these theories have yet to make a significant impact on travel modeling practice. This thesis presents a unifying interdisciplinary framework for a theory of travel decision-making with applications for travel demand modeling and forecasting and a focus on walking and bicycling. The framework offers a guide for future research examining the complex relationships of activities, built environment factors, demographic and socioeconomic characteristics, attitudes and perceptions, and habit and exploration on individual short-term travel decisions (with considerations of the influence of medium- and long-term travel-related decisions). A key component of the theory is a hierarchy of travel needs hypothesized to be considered by travelers in the course of their decision-making processes. Although developed to account for the factors that particularly influence decisions surrounding walking and bicycling, the framework is postulated to apply to all travel modes and decisions, including frequency, destination, mode, time-of-day, and route. The first section of the thesis reviews theories from the fields of economics, geography, psychology, and travel behavior that have a large influence on the development of the theory of travel decision-making. In the next and largest chapter, the components and relationships in this theory, including the hierarchy of travel needs, are defined and presented with supporting empirical evidence from travel behavior research. This thesis's final section views the theory of travel decision-making through the lens of applicability to travel demand modeling and forecasting. The state of current travel forecasting tools, travel behavior research, data, and analysis methods with respect to each aspect of the theory is reviewed. Research and data needs are identified. In closing, some opportunities for operationalizing the theory in travel demand models and using these transportation planning tools for analyzing walking, cycling, and other policies are hypothesized and discussed. This thesis, and the theory and applications discussed within, contribute to the academic study of travel behavior, the practical modeling of travel demand, and walking and bicycling research and planning.

University of Macau / Faculty of Science and Technology / Department of Mathematics

Thesis (M.S.)--Florida State University, 2004. / Advisor: Dr. T.N. Krishnamurti, Florida State University, College of Arts and Sciences, Dept. of Meteorology. Title and description from dissertation home page (Jan. 20, 2005). Includes bibliographical references.

Thesis (M.B.A.)--University of Hong Kong, 1996. / Includes bibliographical references (leaf 67-68).

En effektiv materialhantering har en viktig del i hur konkurrenskraftig en organisation är och att utveckla sina arbetssätt inom detta område är en kontinuerlig utmaning. I en organisation utgör lagerhållning av varor en av de mest väsentliga logistikkostnaderna. Ett sätt att reducera dessa kostnader är genom prognostisering.  Det finns därmed ett incitament för organisationer att allokera resurser på detta område för att bedriva sin verksamhet mer kostnadseffektivt och vara fortsatt konkurrenskraftiga. Statliga myndigheter är inget undantag gällande kostnadsreduceringar, speciellt eftersom de hushåller med resurser som tillhandahålls av regeringen. Studiens syfte har varit att undersöka vilka kvantitativa prognosmetoder som kan användas och hur de ska väljas för att uppnå en förbättrad lagerstyrning. Studien genomfördes som en fallstudie på Trafikverket. Genom att studera en organisation och använda verklig data, bidrar studien med en ökad förståelse för hur prognostiseringsprocessen är utformad samt hur prognostisering kan tillämpas utifrån efterfrågan på ett brett sortiment av artiklar. Totalt har prognoser upprättats för 245 artiklar som finns i det nuvarande sortimentet. Fem olika prognosmetoder användes för att upprätta prognoser som sedan utvärderades med två prognosmått. För att sedan välja den bästa prognosmetoden för varje artikel på ett objektivt sätt användes Data Envelopment Analysis (DEA). Studien visar vilka steg som ingår i prognostiseringsprocessen samt vilka aktiviteter som bör genomföras i varje steg. Studiens användning av DEA-analys för att bestämma den mest lämpliga prognosmetoden för ett sortiment av artiklar visar nya användningsområden för metoden. Studiens resultat och analys visade att de prognosmetoder som organisationen använder inte genererar det bästa resultatet. Prognosfelet minskade för sortimentet när den prognosmetod som DEA-analysen föreslog användes. Studiens slutsatser visade att organisationen inte analyserat varför de prognosmetoder som används är lämpliga samtidigt som de inte följer upp och utvärderar resultatet. Vidare gavs rekommendationer till fallstudieföretaget gällande vilka prognosmetoder och utvärderingsmått de kan använda sig av. / For organizations, an efficient management of material is a source for cost reductions and is an important aspect to continuously develop in order to stay competitive. Keeping stock is one of the significant costs related to logistics and one method that can contribute to a reduction in future stock levels is demand forecasting. There is therefore an incentive for organizations to allocate resources to this area of subject, to operate more efficiently and maintain or increase competitiveness. The public sector is not an exception in regard to working with cost reductions and a strict budget, especially since they use funds allocated by the government. The purpose of this thesis is to investigate which quantitative forecasting methods that can be used and how the methods should be chosen. The goal is to provide more accurate forecasts that contributes to a reduction of total material handled. The thesis was performed as a case study at Trafikverket (the Swedish transport administration). Through study and analysis of a real organization who provided real-world data to work with, the thesis contributes with an understanding of how the forecasting process is outlined and how it can be practiced when managing a large set of products. Five different forecasting methods were used in this thesis together with two measurements of accuracy. In total, forecasts were developed for 245 different products by using real data on past demand. Data Envelopment Analysis (DEA) was used to choose the most and least appropriate forecasting methods for each product. The thesis demonstrates which steps and activities are necessary in the forecasting process and the usage of DEA to decide which method is the most and least suitable for a large set of products shows a new usage for the method. The result and analysis of the thesis showed that the forecasting methods used by the case company are not the most suitable ones. The forecast error was reduced for the total set of products when using the methods that the DEA suggested. The conclusions of the thesis revealed that the case company did not perform any analysis on why the current forecasting methods were used and adding to this, no evaluation of the forecast error was performed. Further, recommendations were provided to the case company regarding which forecasting methods and accuracy measures should be employed.

Trend forecasting is perceived to be an essential service for fashion companies to use in order to stay competitive in the fast-paced fashion industry. Yet, in times of climate change, appointing new trends each season is a questioned practice. Since trend forecasting aligns with the inherent obsolescence of fashion’s constant change, forecasting seems to stand in paradox with the imperatives of sustainability. Thus, this thesis aims to explore the role of trend forecasting to understand its compatibility with environmental sustainability. The review of previous research depicts the evolution of the trend forecasting field and displays prominent literature within fashion and sustainability, which together displays an apparent research gap that this study aims to fill. The thesis follows an exploratory design pursuing a multiple case study strategy applied through eight semi-structured interviews with trend forecasters and a content analysis of WGSN online trend forecasts. Findings of this study validate the existence of a trend forecasting paradox while also demonstrating areas of compatibilities. Customized forecasting and long-term approaches were concluded as compatible practices and may be integrated into both long-term and seasonal forecasting. This study also recognizes a need to differ between forecasting sustainability and sustainable forecasting. This thesis is believed contribute to an under-researched area and aid the trend forecasting industry to realize its impact on sustainability, as well as suggesting approaches on how to further incorporate sustainable practices into their work.

Reliability and accuracy of the forcing data plays a vital role in the Hydrological Streamflow Prediction. Reliability of the forcing data leads to accurate predictions and ultimately reduction of uncertainty. Currently, Numerical Weather Prediction (NWP) models are developing ensemble forecasts for various temporal and spatial scales. However, it is proven that the raw products of the NWP models may be biased at the basin scale; unlike model grid scale, depending on the size of the catchment. Due to the large space-time variability of precipitation, bias-correcting the ensemble forecasts has proven to be a challenging task. In recent years, Ensemble Pre-Processing (EPP), a statistical approach, has proven to be helpful in reduction of bias and generation of reliable forecast. The procedure is based on the bivariate probability distribution between observation and single-value precipitation forecasts. In the current work, we have applied and evaluated a Bayesian approach, based on the Copula density functions, to develop an ensemble precipitation forecasts from the conditional distribution of the single-value precipitation. Copula functions are the multivariate joint distribution of univariate marginal distributions and are capable of modeling the joint distribution of two variables with any level of correlation and dependency. The advantage of using Copulas, amongst others, includes its capability of modeling the joint distribution independent of the type of marginal distribution. In the present study, we have evaluated the capability of copula-based functions in EPP and comparison is made against an existing and commonly used procedure for same i.e. meta-Gaussian distribution. Monthly precipitation forecast from Climate Forecast System (CFS) and gridded observation from Parameter-elevation Relationships on Independent Slopes Model (PRISM) have been utilized to create ensemble pre-processed precipitation over three sub-basins in the western USA at 0.5-degree spatial resolution. The comparison has been made using both deterministic and probabilistic frameworks of evaluation. Across all the sub-basins and evaluation techniques, copula-based technique shows more reliability and robustness as compared to the meta-Gaussian approach.

Every new product launching industrial company faces the difficulties of forecasting future success or failure of a new product before launch. Before launch it is common to develop a business case in order to estimate future quantities and set prices. In the present paper the challenges of developing a standardized business case tool for a large industrial construction and mining company are presented. Few academic studies have been conducted on the challenges and complexities of developing business cases. The research question under which this study is done is: What are the challenges associated with developing an effective standardized business case tool for a large industrial construction and mining company? Due to the different subject areas of the business case for new product launch, the challenges are categorized by topics developed by the researcher in the course of this project: process and team, data gathering and validation, quantity forecast and price forecast. The main challenges found in these categories by the researcher are: finding and motivating experts for the project of developing a standardized business case, gathering and selecting all data necessary without including redundant data, ensuring that different potential new products can be forecasted and designing the price forecast to be profit-maximizing. Solutions to these challenges are provided in the context of a case company by using methods suggested by the academic literature and the evaluation of expert interviews inside the case company

No description available.