Abstract
M.Tech. (Industrial Engineering)
The application of effective demand management practices is important in achieving the profit motives and sustainability of a business. Demand management practice aids businesses in being proactive towards anticipated demands, reactive towards unanticipated demands, and to identify deficiencies that prevent them from achieving customer objectives. Forecasting, as an integrated part of effective demand management, plays a fundamental role in predicting future customer demands for empty wagons.
The purpose of this study is two-fold: firstly, to analyze the current demand management practices employed by the case company, primarily focusing on forecasting activities concerned with the distribution of empty wagons; and secondly, to develop and recommend a forecasting technique for the prediction of empty wagons.
In achieving the purpose of the research study, a case study research design was conducted applying a deductive approach and mixed method research strategy. Furthermore, in evaluating the proposed time series method, historical data of actualized tonnages of fuel distributed by empty wagons covered a four year period from April 2012 to March 2016. The data was divided into two parts: the in-sample data was used to fit the method, and the out-of-sample data was employed to generate forecasts.
The analysis highlighted deficiencies of the current demand management practices of the case company; and highlighted that another integral part of the demand management practices is currently applied internally by the case company. Although the case company does not generate its own forecasts internally using a quantitative forecasting approach, it uses a qualitative forecasting approach that is highly dependent on the supplied forecasts by its customers. Furthermore, the recommended time series method yielded acceptable results when applied to the in-sample data based on the Root Mean Squared Error (RMSE), Mean Absolute Error (MAE) and Mean Absolute Percentage Error (MAPE) values. However, the out-of-sample data for the forecasting...