Demand Forecasting Techniques and Exercises
Uploaded by
ScribdTranslationsDemand Forecasting Techniques and Exercises
Uploaded by
ScribdTranslationsFORECASTING EXERCISES
1. Moving average and exponential smoothing
The historical demand for sales in the last months of a product is:
MES QUANTITY
(tons)
September 12
October 14
November 14
December 15
January 18
February 22
a) Using the weighted moving average method with weights 0.6, 0.3, and 0.1, find the
forecast for the month of March.
b) Using the simple moving average for 4 months, find the forecast for March
c) Using exponential smoothing with α = 0.2 and a forecast for December of 16 tons,
determine what the forecast for January was.
d) Using exponential smoothing with α = 0.2 and forecast for February of 15 tons,
determine the forecast for March.
2. Demand forecast: simple rule (focused projection)
The following table presents the sales (in units) of a product.
MES 2014 2015
January 1760 1810
February 1815 1860
March 1870 1920
April 1925 1870
May 2090 1850
June 1980 1930
July 1760 2000
August 1760 2050
September 1870 2050
October 1540 2080
November 1650 2120
December 1760 2150
Using each of the following simple forecasting rules, determine the expected demand for the
first quarter of the year 2016.
Forecasting rules:
The total sales for the next quarter will be 90% of the total sales of the previous quarter.
The sales for each of the months of the next quarter will be equal to what was sold in each of the
months of the same quarter of the previous year.
What will be sold in each of the next three months will be 5% more than what was sold in each
from the months of the previous quarter.
What will be sold in total in the next three months will be 5% less than what was sold in total the same
quarter of the previous year.
What will be sold in the next three months will be equal to the average of what was sold in the two quarters.
previous.
The rate of change between the previous quarter and the same quarter of the previous year is equal to the rate of change
expected between the next quarter and its equivalent of the previous year.
1
3. Demand forecast: moving averages Month Demand
For a cake factory whose historical demand for fresh cakes in
March 10
thousands of dozens is:
April 18
May 26
You are asked:
June 25
a) Use the simple three-month moving average to calculate the
prediction for the month of August. July 26
b) Use a weighted moving average with weights of 0.6, 0.3, and 0.1 to find the prediction.
of the month of August.
4. Demand forecast: trend
Prepare the quarterly forecast of hours reserved (in thousands)
Trimestre 2007 2008 2009
for the fourth quarter of the year 2010 based on the information
I 17 24
next:
II 20 25
III 30 22 28
IV 17 20
5. Demand forecast: trend
The historical monthly demand for a product expressed in units is:
MES CANT.
October 12
November 15
December 15
January 18
February 22
March 25
Using a linear regression analysis, determine the mathematical model that represents the demand.
of the product and calculate the forecast for the upcoming months of July and December.
6. Demand forecast: trend
The company DESHIDRATADORA DEL PACIFICO S.A. wants to explore the
possibility of opening a new flour production plant SALES
Lucuma that replaces the current one. So that the location decision Year (Ton.)
the new plant being a good long-term decision, the 2005 25
the general management of the company has decided to locate the new 2006 26
plant in a place that allows for economical production 2007 28
demand for the next ten years. For this, it has provided him with 2008 30
the company's sales over the last five years. This 2009 32
information appears below:
The general management is asked to assist in calculating the sales forecasts for the years 2015 and 2019.
7. Forecast with linear regression
Carmela Benavente, Director of Pediatric Primary Care at the Ministry of Health, is evaluating
the current reality condition of its target care segment, which includes children from 0 (zero) to
14 (fourteen) years in all social strata of the country. To this end, it has consulted with INEI.
about the population in that age group at the national level. They have provided you with the following details:
htp://[Link]/esstatistics/index-themes/population-and-housing/
The team led by Dr. Benavente has determined that it will conduct an analysis of the situation due to
every FIVE YEARS for the total population of CHILDREN AGED ZERO TO FOURTEEN every June 30.
Hello has prepared the following summary table:
2
2,000 2.005 2.010 2.015
0-4 3.004.670 2,983,020 2.958.307 2.861.874
5-9 2.981.112 2,959,627 2.938.148 2.922.744
10-14 2,880,364 2.950.813 2.926.874 2.914.162
Total 8,866,146 8.893.460 8,823,329 8,698,780
Use the data from the summary table, only for the TOTAL population every 5 years, and generate the model of
Least Squares Linear Regression, with the projection for the years 2020 and 2025. Show all the
necessary columns for the calculation, with two decimals. What does MADtene mean?
8. Forecast with linear regression
The company ACME has a subsidiary in the capital of a country in Asia. The information it has is
the following:
Population of GDP country (thousands Sale (in
Id Year city (in thousands) millions $ miles)
1 2015 1,505 201 504
2 2016 1,650 208 522
3 2017 2,050 230 560
4 2018 1,680 250 540
5 2019 1,701 280 549
Determine what the sales forecast is for the year 2022.
9. Linear regression with seasonality and exponential smoothing
We have the following demand history for boats in a factory:
Year
Season 2004 2005 2006 2007
Winter 1400 1200 1000 900
Spring 1500 1400 1600 1500
Summer 1000 2100 2000 1900
Autumn 600 750 650 500
a) Use the deseasonalized linear regression model to make the sales forecast
for the winter of 2009.
b) Apply the exponential smoothing model with trend (Holt) to make the autumn projection.
from 2008, with values Alpha = 0.1 and Beta = 0.2
c) Apply the exponential smoothing model with trend and seasonality (Winter), with Alpha = 0.1,
Beta = 0.2 y Gamma = 0.1
d) Graphically compare the previous forecasts and indicate which methodology fits best with the
data Why?