This action might not be possible to undo. Are you sure you want to continue?

# BY ²ISHANI DWIVEDI RICHA THAKUR

INTRODUCTIONy This is the method of demand forecasting. y It tries to overcome the limitation of moving y y y y y

averages and eliminates the records of passed data. It also tries to screened out the irregularities in the demand pattern. This method takes into consideration the short term fluctuations. It decreases in geometric progression as we move towards the old observations. In this method the later figures reflect more up to date average of the series. This is the easier method of achieving a progressing from of weighting.

**CONCEPT OF EXPONENTIAL SMOOTHING
**

y New estimate =old estimate of latest actual demand+ (latest actual demand-old estimate of latest demand) y Formulay y Where,

Ft= .Dt+(1- )Ft-1= Ft-1+ et

y Ft is the forecast at time t. y Dt is the actual demand at time t. y Ft-1 is the forecast at time (t-1) y

is the smoothing coefficient, et=(Dt-Ft-1)

**METHOD OF FINDING ERROR IN THE FORECAST OF DEMAND
**

y First of all find error by subtracting the recent average from the latest in coming observation. y Multiply error et by . y Add .et to the Ft. this gives new Ft as the forecast the next period.

PERFORMANCE

y The performance of this method depends on the value of smoothing coefficient and the initial forecast Ft-1.

gives the selector a major of control over the degree of smoothing induced in the series. y The choice of depends upon how much weight is desired to be given to later periods relative to earlier periods. y Low value of gives more weightage to the past figure and are used where the series is rather stable and vice versa.

y

y The value of

practice and 0. . y In cyclical fluctuations we take low value of and in case of long term fluctuations we take high value of .

lies between 0 and 1 and in is generally taken between 0.1

EXPONENTIAL

SMOOTHING FORECASTING SYSTEM

Initial Ft-1 demand

Transformation Process Ft-1+ et Feedback et

Forecas t

Ft

ADVANTAGES

y Exponential smoothing provides a

convenient systematic method . y The amount of adjustment is determined by the selected smoothing coefficient the greater the coefficient the greater is the adjustment and vice-versa.

EXAMPLEy Forecast the demand for the following

series by exponential smoothing method: y Period (t) 1 2 3 4 5 6 7 8 9 10 11 12 y Actual Demand (Dt) 10 12 8 11 9 10 15 14 16 15 14 15 SOLUTION- The forecasts for various periods can be calculated in the following tabular form. Here we consider =0.1 and then compare the forecasts for =0.7.

THE INITIAL FORECAST IS TAKEN TO BE 10 FOR PERIOD:Period Actual Ft-1 demand =0.1 Ft 10.00 10 12 8 11 9 10 15 14 16 15 14 15 10.00 10.00 10.20 9.98 10.08 9.97 9.97 10.47 10.82 11.34 11.71 11.94 10.00 10.20 9.98 10.08 9.97 9.97 10.47 10.82 11.34 11.71 11.94 12.25 0.0 2.00 -2.20 1.02 -1.08 0.03 5.03 3.53 5.18 3.66 2.29 3.06 10 10 11.4 9.02 10.41 9.42 9.83 13.45 13.81 15.35 15.11 14.33 et Ft-1 =0.7 Ft 10 10 11.4 9.02 10.41 9.42 9.83 13.45 13.81 15.35 15.11 14.33 14.80 0.0 2.0 -3.40 1.98 -1.41 0.58 5.17 0.55 2.16 -0.35 -1.11 0.67 et

0 1 2 3 4 5 6 7 8 9 10 11 12

y Now we calculate MDA= y

| et | /1 2 =2.42 for =0.1

and 1.62 for

=0.7 since MAD for =0.7 is lesser than MAD for =0.1, =0.7 gives better forecast.

BOX ± JENKINS METHODy This forecasting approach has been

borrowed from the control theory. y In a control system there is a target which has to be tracked where the target itself is moving or changing its value or position. y The control mechanism measures the difference between the target and actual output. y Any error is used along with its derivatives and integrals for improving the forecast for the next period and so on.

y The basic equation is:y New forecast =old forecast+the direct

term:A(error in old forecast)+time derivative term:B(change in error between last time and time immediately before)+the integral term: C(sum of the errors so far) y Ft=Ft-1 +(r-1)(et-1 ²et-2)+re(et-1)+r1 et

THANK YOU