2015 International Conference on Industrial Engineering and Operations Management (IEOM), Dubai, United Arab Emirates, 3 - 05 March 2015
Forecasting and Time series techniques are frequently used and play extremely important roles in managerial activities and decision-making processes. Holt-Winters (HW) which is one of the most popular forecasting methods is utilized in cases where data show seasonality and/or trend. The method involves the selection of several parameters for optimum prediction results. Heuristics are usually utilized for optimum parameter selection and it appears to be an open field for improvement. In this study "Spreadsheet modeling" of HW method is improved by minimizing the mean squared error (i.e. prediction error). Since the forecast error is nonlinear function; Holt-Winters parameter optimization in this study is achieved by "Excel Nonlinear Solver" and "Differential Evolution Search" techniques. To increase the usability of the spreadsheet modeling advance macro programming techniques are incorporated with the developed models in Microsoft Excel.