FORECASTING OF WINTER’S EXPONENTIAL SMOOTHING USING NON-LINEAR PROGRAMMING ALGORITHM

Abstract

Forecasting of Winter’s exponential smoothing (WES) multiplicative can be used, when value of parameter level ( ), trend ( ), and seasonal ( ) exist with little error. Generally , , and are found by trial and error method. Non-linear programming algorithm can used for getting , , and . This thesis concerns in quadratic algorithm to get , , for WES. The values of , , found by quadratic algorithm using software SPSS are , , and . Respectively in the simulation (or applied) on the read data, the results of forecasting are 386,901,811,018.13 rupiahs, 393,361,659,799.84 rupiahs, 399,335,147,151.67 rupiahs, 404,859,870,667.20 rupiahs, and 409,970,521,748.95 rupiahs. Forecasting income of Tegal for 5 next year can be expressed optimum, caused by , , have little forecasting error with MAPE is 72.7360.

Key Words : forecasting, WES multiplicative, non-linear programming