Problems of Forecasting the demand

Problem 3

Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec

Last 100 125 135 175 185 200 150 140 130 200 225 250
year

This 125 135 135 190 200 190

year

To forecast the demand of the third quarter of this year using the three quarter moving
average method we’re gonna add the values of the actual three last month and divided by
the number of month

Jul = 190+200+190/3 = 193.3

For the other months same process but we’re gonna take the value we find for July
Aug = 200+190+193.3/3 = 194.3
Sep = 190+193.3+194.3/3 = 192,52

So with the three quarter moving average method we can say that for July, August and
September the forecasting demand will be 193.3, 194.3 and 192.52

Problem 11

a) If we use the three weeks moving average method we have to take the values of the
3 previous weeks and divide by the number of weeks.
So (400+600+700)/3 = 567
The forecasting demand for the next week should be approximately 567.

b) Initial forecast = 350 (300+400/2)

Alpha = 0.2

Exponial Actual Forecast MAD (A-F)

smoothing (week)

Average Week 1 350 350 (initial

and 2 forecast)

Week 3 600 350+(0.2*(600-350 200

) = 400

Week 4 700 400 240

+(0.2*(700-400) =
460
 Week 5 567 = 85.6
460+(0.2*(567-460
) = 481.4

Problem 13

a) To forecast October we take the average of the last 4 months

= June + July + August + September
= 75 + 60 + 80 + 75 / 4
= 72.5

b)
Actual Exponential smoothing forecast

Septembre = 75 Initial Forecast = 65

October = 65 +((0.2*(75-65)) = 67
The forecast for October will be 67.

c)
Demand (Y) Time (X) Time square Y*X
(X²)

April (60) 1 1 60

May (55) 2 4 110

June (75) 3 9 150

July (60) 4 16 240

August (80) 5 25 400

September 6 36 450
(75)

Total (405) Total (21) (91) Total (1410)

Average 67.5 3.5 15.17 235

b = Sum (Y)*SumX² - Sum X*Sum(X*Y)/n*Sum(X²)-Sum(X)²

= (405*91)-(21*1410)/(6*91) - (21)²
= 7245/105
= 69

a = ( n *sum (XY) - (SumX)*SumY)))/N*SumX²*Sum X)²

= 6*1410 - (21*405)/(6*91 - 21²)
= -45/105
= -3/7

Trend equation T = a +(b*X)

= 3/7X + 69

d) October is gonna be the 7 month

So - 3/7*7 + 69
= 66

The forecast for October using the regression formula is 66.

Problem 17

Month Actual

January 110

February 130

March 150

April 170

May 160

June 180

July 140

August 130

September 140

a) Using the three month moving average

Forecast for April

- =150+170+160)/3=160
Forecast for May:
- (130+150+170)/3=150
Forecast for June:
- (150+170+160)/3=160

Forecast for July:

- (170+160+180)/3=170
Forecast for August:
- (160+180+140)/3=160

Forecast for September:

- (180+140+130)/3=150

b) Exponential smoothing method

​ Initial Forecast for April (using the average of January through March):

(110+130+150)/3=130

Forecast for April = alpha*Actual April + (1-alpha)

0.3×170+(1−0.3)×130
0.3×170+0.7×130=51+91=142

Forecast for May =

0.3×160+(1−0.3)×142
0.3×160+0.7×142=48+99.4≈147
Forecast for June =
0.3×180+(1−0.3)×147
0.3×180+0.7×147=54+102.9≈157

For July, August, and September, continue the same process using the updated forecasts:

July Forecast ≈ 164

August Forecast ≈ 158
September Forecast ≈ 149

c) Which method choose

To decide which method produced the better forecast over the six-month period, we can use
the Mean Absolute Deviation (MAD). MAD measures the average magnitude of errors
between forecasts and actual values. The method with the lower MAD is considered to have
produced the better forecast.

Let's calculate the MAD for both methods:

​ Three-Month Moving Average:
Actual Demand:
April: 170
May: 160
June: 180
July: 140
August: 130
September: 140

Forecasted Demand (from the previous calculation):

April: 160
May: 150
June: 160
July: 170
August: 160
September: 150

Absolute Errors:
April: |170 - 160| = 10
May: |160 - 150| = 10
June: |180 - 160| = 20
July: |140 - 170| = 30
August: |130 - 160| = 30
September: |140 - 150| = 10

MAD for Three-Month Moving Average =

(10+10+20+30+30+10)/6 ≈18.33

Simple Exponential Smoothing with Alpha of 0.3:

Absolute Errors (from the previous calculation):

April: |170 - 142| = 28

May: |160 - 147| = 13
June: |180 - 157| = 23
July: |140 - 164| = 24
August: |130 - 158| = 28
September: |140 - 149| = 9
MAD for Simple Exponential Smoothing =
(28+13+23+24+28+9)/ 6 ≈20.83

Comparing the MAD values, we find that the Three-Month Moving Average method has a
lower MAD of approximately 18.33, while the Simple Exponential Smoothing method has a
MAD of approximately 20.83. Therefore, based on the MAD criterion, the Three-Month
Moving Average method produced the better forecast over the six-month period.

Problem 20 :

Week Forecas Actual Deviatio RSFE Absolut Sum of MAD Trackin

t n e absolut g signal
deviatio e
n deviatio
n

1 800 900 100 100 100 100 100 1

2 850 1000 150 250 150 250 125 2

 3 950 1050 100 350 100 350 117 3

4 950 900 -50 300 50 400 100 3

5 1000 900 100 200 100 500 100 2

6 975 1100 125 325 125 625 104 3.1

1) Cumulative sum of the forecast eros = 100+150+100+50+100+125

= 625
MAD = 625/6
= 104.17

For week number 6 the MAD is 104 and the tracking signal is 3.1 which is a lot. This
forecasting model is unacceptable because the value is too high.

Problem 25

Months Sales Months Sales

January-February 109 January-February 115

March-April 104 March-April 112

May-June 150 May-June 159

July-August 170 July-August 182

September-October 120 September-October 126

November-Decemb 100 November-Decemb 106

er er

a) Plot the data

 b) Let’s calculette the necessary sums
Period = 1+2+3+4+5+6+4+5+6+7+8+9+10+11+12
= 78

Sales = 109+104 + 150 ……+ 106 = 1553

XY = 109, 208, 450, 680, …..1272

= 10257

X² =1² + 2² + 3² + 4² …. + 12²
= 650

a = (12*10257) - (78*1553)/(12*650) - (78)²

= 1950/7800 - 6084
= 0.88

b= (1553*650) - (78*10257)/ (12*650) - (78)²

= 209404/1716
= 122.03

Sales = 0.88*Period + 122.03

c) Determine multiplicative seasonal index factor

Calculate the average sales for each period

 Period Sales

1 (109 + 115) / 2 = 112

2 (104 + 112) / 2 = 108

3 (150 + 159) / 2 = 154.5

4 (170 + 182) / 2 = 176

5 (120 + 126) / 2 = 123

6 (100 + 106) / 2 = 103

​ Calculate the average sales for each month:

Month Average Sales

Jan (112 + 108) / 2 = 110

Feb (112 + 108) / 2 = 110

Mar (154.5 + 154.5) / 2 = 154.5

Apr (154.5 + 154.5) / 2 = 154.5

May (123 + 123) / 2 = 123

Jun (123 + 123) / 2 = 123

Jul (176 + 176) / 2 = 176

Aug (176 + 176) / 2 = 176

Sep (103 + 103) / 2 = 103

Oct (103 + 103) / 2 = 103

Nov (110 + 110) / 2 = 110

Dec (110 + 110) / 2 = 110

January:
Average Sales/Overall Average Sales=110/130=0.8462

February:
Average Sales/Overall Average Sales=110/130=0.8462

March:
Average Sales/Overall Average Sales=154.5/130=1.1885

April:
Average Sales/Overall Average Sales=154.5/130=1.1885

May:
Average Sales/Overall Average Sales=123/130=0.9462

June:
Average Sales/Overall Average Sales=123/130=0.9462

July:
Average Sales/Overall Average Sales=176/130=1.3538

August:
Average Sales/Overall Average Sales=176/130=1.3538

September:
Average Sales/Overall Average Sales=103/130=0.7923

October:
Average Sales/Overall Average Sales=103/130=0.7923

November:
Average Sales/Overall Average Sales=110/130=0.8462

December:
Average Sales/Overall Average Sales=110/130=0.8462

d) Forecast sales of next year

January Forecast=(0.88×7+122.03)×0.8462
≈106.781
February Forecast=(0.88×8+122.03)×0.8462
≈108.141

March Forecast=(0.88×9+122.03)×1.1885
≈154.298

April Forecast=(0.88×10+122.03)×1.1885
≈156.483

May Forecast=(0.88×11+122.03)×0.9462
≈126.507

June Forecast=(0.88×12+122.03)×0.9462
≈128.155

July Forecast=(0.88×13+122.03)×1.3538
≈186.999

August Forecast=(0.88×14+122.03)×1.3538
≈189.535

September Forecast=(0.88×15+122.03)×0.7923
≈112.396

October Forecast=(0.88×16+122.03)×0.7923
≈113.904

November Forecast=(0.88×17+122.03)×0.8462
≈123.138

December Forecast=(0.88×18+122.03)×0.8462
≈124.680

Months Sales forecast next year

January 106.781

February 108.141

March 154.298

April 156.483

May 126.507

June 128.155
July 186.999

August 189.535

September 112.396

October 113.904

November 123.138

December 124.680