# Forecasting. Lecture Outline Strategic Role of Forecasting in Supply Chain Management and...

date post

14-Jan-2016Category

## Documents

view

216download

1

Embed Size (px)

### Transcript of Forecasting. Lecture Outline Strategic Role of Forecasting in Supply Chain Management and...

Forecasting

Lecture OutlineStrategic Role of Forecasting in Supply Chain Management and TQMComponents of Forecasting DemandTime Series MethodsForecast AccuracyRegression Methods

ForecastingPredicting the FutureQualitative forecast methodssubjectiveQuantitative forecast methodsbased on mathematical formulas

Forecasting and Supply Chain ManagementAccurate forecasting determines how much inventory a company must keep at various points along its supply chainContinuous replenishmentsupplier and customer share continuously updated datatypically managed by the supplierreduces inventory for the companyspeeds customer deliveryVariations of continuous replenishmentquick responseJIT (just-in-time)VMI (vendor-managed inventory)stockless inventory

Forecasting and TQMAccurate forecasting customer demand is a key to providing good quality serviceContinuous replenishment and JIT complement TQMeliminates the need for buffer inventory, which, in turn, reduces both waste and inventory costs, a primary goal of TQMsmoothes process flow with no defective itemsmeets expectations about on-time delivery, which is perceived as good-quality service

Types of Forecasting MethodsDepend ontime framedemand behaviorcauses of behavior

Time FrameIndicates how far into the future is forecastShort- to mid-range forecasttypically encompasses the immediate futuredaily up to two yearsLong-range forecastusually encompasses a period of time longer than two years

Demand BehaviorTrenda gradual, long-term up or down movement of demandRandom variationsmovements in demand that do not follow a patternCyclean up-and-down repetitive movement in demandSeasonal patternan up-and-down repetitive movement in demand occurring periodically

Forms of Forecast Movement

Forecasting MethodsQualitativeuse management judgment, expertise, and opinion to predict future demandTime seriesstatistical techniques that use historical demand data to predict future demandRegression methodsattempt to develop a mathematical relationship between demand and factors that cause its behavior

Qualitative MethodsManagement, marketing, purchasing, and engineering are sources for internal qualitative forecastsDelphi methodinvolves soliciting forecasts about technological advances from experts

Forecasting ProcessNoYes

Time SeriesAssume that what has occurred in the past will continue to occur in the futureRelate the forecast to only one factor - timeIncludemoving averageexponential smoothinglinear trend line

Moving AverageNaive forecastdemand the current period is used as next periods forecastSimple moving averagestable demand with no pronounced behavioral patternsWeighted moving averageweights are assigned to most recent data

Moving Average:Nave Approach

Simple Moving Average

3-month Simple Moving Average

5-month Simple Moving Average

Smoothing Effects

Weighted Moving AverageAdjusts moving average method to more closely reflect data fluctuations

Weighted Moving Average ExampleMONTH WEIGHT DATA

August 17%130September 33%110October 50%90

Exponential SmoothingAveraging method Weights most recent data more stronglyReacts more to recent changesWidely used, accurate method

Exponential Smoothing (cont.)Ft +1 = Dt + (1 - )Ftwhere:Ft +1 =forecast for next periodDt =actual demand for present periodFt =previously determined forecast for present period=weighting factor, smoothing constant

Effect of Smoothing Constant0.0 1.0 If = 0.20, then Ft +1 = 0.20Dt + 0.80 FtIf = 0, then Ft +1 = 0Dt + 1 Ft 0 = Ft Forecast does not reflect recent dataIf = 1, then Ft +1 = 1Dt + 0 Ft =Dt Forecast based only on most recent data

Exponential Smoothing (=0.30)

Exponential Smoothing (cont.)

Exponential Smoothing (cont.)

Adjusted Exponential SmoothingAFt +1= Ft +1 + Tt +1whereT = an exponentially smoothed trend factor

Tt +1 = (Ft +1 - Ft) + (1 - ) TtwhereTt = the last period trend factor= a smoothing constant for trend

Adjusted Exponential Smoothing (=0.30)

Adjusted Exponential Smoothing: Example

Adjusted Exponential Smoothing Forecasts

Linear Trend Liney = a + bx

wherea = interceptb = slope of the linex = time periody = forecast for demand for period x

Least Squares Example

Least Squares Example (cont.)

Seasonal AdjustmentsRepetitive increase/ decrease in demandUse seasonal factor to adjust forecast

Seasonal Adjustment (cont.)

Seasonal Adjustment (cont.)

Forecast AccuracyForecast errordifference between forecast and actual demandMADmean absolute deviationMAPDmean absolute percent deviationCumulative errorAverage error or bias

Mean Absolute Deviation (MAD)where t= period number Dt= demand in period t Ft= forecast for period t n= total number of periods= absolute value

MAD Example

Other Accuracy Measures

Comparison of Forecasts

Forecast ControlTracking signalmonitors the forecast to see if it is biased high or low

1 MAD 0.8 Control limits of 2 to 5 MADs are used most frequently

Tracking Signal Values

Tracking Signal Plot3 2 1 0 -1 -2 -3 |||||||||||||0123456789101112Tracking signal (MAD)Period

Statistical Control ChartsUsing we can calculate statistical control limits for the forecast errorControl limits are typically set at 3

Statistical Control Charts

Regression MethodsLinear regressiona mathematical technique that relates a dependent variable to an independent variable in the form of a linear equationCorrelationa measure of the strength of the relationship between independent and dependent variables

Linear Regression

Linear Regression Examplexy(WINS)(ATTENDANCE) xyx2

436.3145.216640.1240.636641.2247.236853.0424.064644.0264.036745.6319.249539.0195.025747.5332.549

49346.72167.7311

Linear Regression Example (cont.)

Linear Regression Example (cont.)

Correlation and Coefficient of DeterminationCorrelation, rMeasure of strength of relationshipVaries between -1.00 and +1.00Coefficient of determination, r2Percentage of variation in dependent variable resulting from changes in the independent variable

Computing Correlation