ME 5204-Production Planning and Control: L2 - Quantitative Forecasting Techniques

You might also like

Download as pptx, pdf, or txt
Download as pptx, pdf, or txt
You are on page 1of 27

ME 5204- Production Planning and Control

L2 – Quantitative Forecasting Techniques

Eng. (Ms.) Hiruni Wijerathne


Department of Mechanical and Manufacturing Engineering
University of Ruhuna
Quantitative Forecasting Techniques
• Simple Moving Average Method
• Weighted Moving Average Method
• Single Exponential Smoothing Average Method
• Double Exponential Smoothing Average Method
• Linear Regression/Simple Regression Average Method

2 of 27
1. Simple Moving Average Method
• The average of demands from most recent periods is taken as
the demand for next time period.

Q1) Detailed report of sold vehicles in ABC company during the


last year, has denoted in the table. What will be the forecasted
demand for vehicles in next January?
a) Using 3 months average
b) Using 5 months average

3 of 27
Month Demand
January 95
February 100
March 87
April 123
May 90
June 96
July 75
August 78
September 106
October 104
November 89
December 83
January ?

4 of 27
Answer:
3 Months 5 Months
Month Demand
Forecast Forecast
January 95
February 100
 
March 87
April 123 94
May 90
June 96 99
July 75
August 78
September 106
October 104
November 89
December 83
January

5 of 27
Answer:
3 Months 5 Months
Month Demand
Forecast Forecast
January 95
February 100
 
March 87
April 123 94
May 90 103.3
June 96 100 99
July 75 103 99.2
August 78 87 94.2
September 106 83 92.4
October 104 56.3 89
November 89 96 91.8
December 83 99.6 90.4
January 92 92

6 of 27
2. Weighted Moving Average Method
• Unequal weights are assigned to the past demand data while
calculating the demand forecast for next time period.

• Most recent data is assigned the highest weight factor.

7 of 27
Q2) Following table shows the computation for the 3 months
weighted moving average with a weight of 0.5 assign to most
recent value, a weight of 0.3 assign to next most recent value
and weight of 0.2 assigned to the oldest of the demand value
included in the average.

Identify the weighted factors (a,b,c) and calculate the


forecasting value for the upcoming month.

8 of 27
Month Demand
01 120

02 130

03 110

04 140

05 110

06 130

9 of 27
Answer:
a=0.2, b=0.3, c=0.5

Forecast
Month Demand Value
01 120
02 130
03 110
04 140 118
05 110
06 130
07

10 of 27
Answer:

Forecast
Month Demand Value
01 120
02 130
03 110
04 140 118
05 110 129
06 130 119
07 129

11 of 27
3. Single Exponential Smoothing Average Method

• Weights are assigned in exponential order.

Ft+1 = .Dt + (1 - ).Ft

Ft+1 = Forecasted Value for Next Period


Dt = Actual Demand for Present Period
Ft = Previously determined forecast for present period.
 = Smoothing factor/ weighting factor (0    1)

12 of 27
Q3) Demand for a particular product of ABC company is shown
in the table. Forecast the expected demand for next January by
using the simple exponential smoothing average method.
a) =0.3
b) =0.5

13 of 27
Period Month Demand

01 January 37
02 February 40
03 March 41
04 April 37
05 May 45
06 June 50
07 July 43
08 August 47
09 September 56
10 October 52
11 November 55
12 December 54

14 of 27
Answer:

Period Month Demand (Dt) Forecast (Ft) Forecast


(=0.3) (=0.5)
01 January 37 37
02 February 40 37
03 March 41 37.90

Ft+1 (Feb) =  .Dt + (1 - ).Ft


= 0.3x37 + 0.7x37

= 37

Ft+1 (Mar) = 0.3x40 + 0.7x37

= 37.90

15 of 27
Answer: Forecast Forecast(
Period Month Demand
(=0.3) =0.5)
01 January 37 37 37
02 February 40 37 37
03 March 41 37.9 38.5
04 April 37 38.8 39.8
05 May 45 38.3 38.4
06 June 50 40.3 41.7
07 July 43 43.2 45.9
08 August 47 43.1 44.5
09 September 56 44.3 45.8
10 October 52 47.8 50.9
11 November 55 49.0 51.5
12 December 54 50.8 53.6
13 January 51.8 53.8

16 of 27
4. Double Exponential Smoothing Average Method

• Assumption: Data doesn’t contain seasonal, cyclic, repeating or


irregular variation.
• Applicable only for the linear trend.

17 of 27
AFt+1 = .Dt +(1 - ).Ft+ Error
Ft+1 = .Dt +(1 - ).Ft
Error = Tt+1
Tt+1= β (Ft+1 - Ft)+(1 - β ).Tt

AFt+1 = Ft+1 + Tt+1

β = Smoothing Constant for trend


Tt+1 = Trend Error

18 of 27
Period Month Demand
01 January 37
02 February 40
Q4) = 0.5, β= 0.3 03 March 41
04 April 37
05 May 45
06 June 50
07 July 43
08 August 47
09 September 56
10 October 52
11 November 55
12 December 54
13 January ?

19 of 27
Answer: Period Month Demand Ft+1 Tt+1 AFt+1
(=0.5)
01 January 37 37 0 37
02 February 40 37 0 37
03 March 41 38.5 0.45 38.95
04 April 37 39.8 0.71 40.51
05 May 45 38.4
06 June 50 41.7
07 July 43 45.9
08 August 47 44.5
09 September 56 45.8
10 October 52 50.9
11 November 55 51.5
12 December 54 53.6
13 January 53.8

20 of 27
5. Linear Regression Forecasting Method
• Mathematical model which obtains the ‘line of best fit’
between the dependent variable(usually demand) and
independent variable.
• Also called as “Square Method”

y = a + bx
Ʃy = Na + b.Ʃx
Ʃxy = a.Ʃx + b.Ʃx2

21 of 27
To compute the values of ‘a’ and ‘b’
1. Calculate the deviation ‘x’ for each period and also the sum
of the deviation, Ʃx
2. Find the values of Ʃx2
3. Find the values of Ʃxy
4. Calculate the values of ‘a’ and ‘b’
5. Make the sum of deviation Ʃx=0

22 of 27
23 of 27
Q5) Following data gives the sales of a company for various years.
Fit the straight line and forecast sales for the years of 1998 and
1999?
Year Sales
1989 13
1990 20
1991 20
1992 28
1993 30
1994 32
1995 33
1996 38
1997 43

24 of 27
Answer: Deviation xy
Year Sales (y) (x) x2
1989 13 -4 16 -52
1990 20 -3 9 -60
1991 20 -2 4 -40
1992 28 -1 1 -28
1993 30 0 0 0
1994 32 1 1 32
1995 33 2 4 66
1996 38 3 9 114
1997 43 4 16 172
Ʃx=0 Ʃx2=60 Ʃxy=204

Find the sales of year 1998 and 1999.

25 of 27
Reasons for out of Forecasting
• Changing trend
• Change the appearance of the product
• Politics
• Weather Change
• Promotions

26 of 27
Recap
• Qualitative Forecasting Techniques
• Reasons for out of Forecasting

27 of 27

You might also like