Abstract:
Because of the demand fluctuations and its labor intensive in a warehouse, the managing of an effective picking activity requires an accurate and reliable workforce preparation. The ineffective preparation equates to high labor costs from idle workers or excessive overtime, similar to a case study company. The company provides dedicated logistics operations in a consuming product distribution center, including put-to-store a reverse of a general picking activity in which an operator consolidates many items by repeatedly dropping specific quantities of an individual item into a buffer area corresponding to an ordered store. In addition to the inaccurate forecasting of incoming cartons, the analysis also reveals that the ineffective preparation causes by a simple workforce forecasting model that neglects mixture of items, quantity of pending items, historical workforce, and productivity of operators. As a result, this study proposes and compares forecasting models that incorporate such factors to predict weekly required man-hour. With training dataset, the model comparison reveals that linear regression models which use new forecasting case as a factor are more accurate than time-series models. In addition, the regression model that combines both new forecasting case, Fast move case and Slow move case provides high accuracy with MAPE of 4.47% Therefore, This model is suitable for the workforce planning.