Outline of Previous Lecture Lecture 7Chapter 5- Demand Forecasting Part 1Current Lecture Lecture 11Chapter 5-Demand Forecasting Part 1 con.Chapter 5-Demand Forecasting Part 2SCM 3301 1st EditionOutline of Previous Lecture Lecture 7Chapter 5- Demand Forecasting Part 1The Role of Demand Forecasting- Designed to estimate future demand for planningo Purchasing Decisionso Inventory Decisionso Production Decisions- Important to match supply with demand- Results of increased forecast accuracyo Lower inventorieso Reduced stock-outso Smoother production planso Reduced costso Improved customer serviceForecasting Techniques- Qualitative Forecastingo Based on opinion & intuition.- Quantitative Forecastingo Using mathematical models & historical data- Components of a TIME SERIESo Trend Variationso Cyclical Variationso Seasonal Variationso Random VariationNOTE: real-world dada incorporates and unpredictable combination of all of the above making forecasting difficultSCM 3301 1st EditionCurrent Lecture Lecture 11Chapter 5-Demand Forecasting Part 1 con.Exponential SmoothingSCM 3301 1st Edition- Exponential Smoothing Forecasto a = 0 All previous data is weighted evenlyo a => Determines how heavily recent data is weighedo a = 1 The most recent data point is weighted exclusively Naïve ApproachFormula: Ft+1=Ft+a (At- Ft)Note: Error= (At-Ft)Linear Trend Forecast- Still a Time Series Forecasting Technique- One Variable Across Time- Time Is The Independent Variable (IV)- Fit Line To Data: Least Squares Analysis- Extrapolate Line Into The FutureCause-and-Effect Models Regression- Dependent (DV) and Independent Variables (IV)- Simple Linear Regressiono One Explanatory (IV) Variableo Difference with Linear Trend: Time No Longer The IV- Multiple Regression Forecasto Multiple variables= more complex, expensive modelo Fewer variables= model not as descriptive of real life- All these techniques are taught ino SCM 4330 Business ModelingChapter 5-Demand Forecasting Part 2Forecast ErrorThe formula for forecast error, defined as the difference between actual quantity and the forcast-Forecast Error, e(t)=A(t)-F(t)Where e(t)= forecast error for Period tA(t)= actual demand for Period t F(t)= forecast for Period tForecast Inaccuracy Cost Money (Tracking Signal)Positive Bias: Forecast is lower than Demand => Stock-outsSCM 3301 1st EditionNegative Bias: Forecast is higher than Demand => InventoryMeasuring Forecast AccuracySeveral measures of forecasting accuracy follow-- Mean absolute deviation (MAD)- a MAD of 0 indicates the forecast exactly predicted demand- Mean absolute percentage error (MAPE)- provides a perspective of the true magnitude of the forecast error- Running Sum of Forecast Errors (RSFE)- indicates bias in the forecasts or the tendency of a forecast- Mean squared error (MSE)Forecast Accuracy Example, RSFE, MAD, MSE, & MAPESCM 3301 1st