MODULE 4

FORECASTING

I. Learning Outcomes

At the end of the module, you should be able to:

1. Solve for the forecast using different methods

2. Identify the most accurate forecasting method for a certain condition/s
or given historical data.

II. Introduction

Before planning for the capacity, it’s important to do forecasting for it will determine
the capacity of the business like in terms of staffing level, inventory levels, factory
capacity etc. Forecasting is vital for business decisions in the areas of production,
inventory, personnel and facilities. These areas are the underlying basis of all business
decisions. So this module will help the Industrial Engineering students learn the common
methods used in forecasting and the underlying concepts of forecasting.

III. Lesson

A. DEFINITION OF TERMS

 Forecasting
Process of predicting a future event.

Process of making predictions of the future based on past and present

data and most commonly by analysis of trends.

B. INFLUENCE OF PRODUCT LIFE CYCLE

In the previous module, Product Life Cycle was discussed. As product

passes through life cycle, forecasts are useful in projecting:

1. Staffing levels
2. Inventory levels
3. Factory Capacity

C. FORECASTING APPROACHES

a) Qualitative Methods
 Used when situation is vague and little data exist (e.g. new products or
new technology)
 Involves intuition, experience (e.g. forecasting sales on internet)
Page | 20
 1. Jury of executive opinion. Pool opinions of high-level experts,
sometimes augment by statistical models

2. Delphi Method. Panel of experts, queried iteratively

3. Sales Force composite. Estimates from individual

salespersons are reviewed for reasonableness, the aggregated

4. Consumer Market Survey. Ask the customer

b) Quantitative Methods
 Used when situation is ‘stable’ and historical data exist (e.g.
existing products or current technology)
 Involves mathematical techniques (e.g. forecasting sales of color
televisions)

TIME SERIES FORECASTING

 Set of evenly spaced numerical data. Obtained by observing
response variable at regular time periods.
important.
Forecast based only on past values, no other variables
It assumes that factors influencing past and present
will continue influence in future.
 Is the use of model to predict future based on previously observed
values
 Type of forecast that uses historical data assuming that future will
likely be the past

1. Naïve Method uses a single previous value of a time series as

the basis of a forecast. The naive approach can be used with a
stable series (variations around an average), with seasonal
variations, or with trend.

2. Simple Moving Average (SMA) is similar to the simple

average except that we are not taking an average of all the
data, but are including only n of the most recent periods in
the average.

3. Weighted Moving Average (WMA) is similar to a moving

average, except that it assigns more weight to the most
recent values in a time series.

4. Exponential Smoothing is a sophisticated weighted averaging

method that is still relatively easy to use and understand. Each
new forecast is based on the previous forecast plus a
percentage of the difference between that forecast and the
actual value of the series at that point.
Page | 21
 5. Linear Trend is a time series technique that computes a forecast
with trend by drawing a straight line through a set of data. This
approach is a version of the linear regression technique.

D. SAMPLE PROBLEMS ON TIME-SERIES FORECASTING

Sample Problem 1 (Simple Moving Average). The Wood Depot sells and
delivers wooden furniture to offices, houses and private institutions within 30-
kilometer radius of its warehouse. The furniture shop business is competitive,
and on time delivery plays a huge factor among customers. The manager of the
company expects enough drivers and trucks to deliver customer orders and that
they also have enough inventories in stock. A manager directs their IE to forecast
the number of orders that will occur the next month. From the records of delivery
orders, below data are gathered for the past 10 months, wherein it is desired to
compute 3-month moving average.

MONTH ORDERS
January 120
February 90
March 100
April 75
May 110
June 50
July 75
August 130
September 110
October 90

Solution

Get the 3 most recent data (August, September and October) and get their
average. 130 + 110 + 90
= = 110 ( )

To find the forecast for December, use the data of September, October,
November (the most recent). Use the actual order for November, not the
forecast value. That’s the concept why it is called moving average.

Sample Problem 2 (Weighted Moving Average). In reference to problem no. 1,

Wood Depot wants to compute a 3-month weighted moving average with a weight of
50 percent for the October data, a weight of 30 percent for the September data, and
a weight of 20 percent for the August data. These weights reflect the
company’s desire to have the closest data point influence the forecast more.

Page | 22
 Solution

Formula for Weighted Moving Average

∑( )
=
∑

= (0.20)(130) + (0.30)(110) + (0.50)(90)

(0.20) + (0.30) + (0.50)

= 104

To find the forecast for December, use the data of September, October,
November (the most recent). Use the same weights (ie. Assign 20% to
October; 30% to October; 50% to November). Use the actual order for
November, not the forecast value. That’s the concept why it is called
weighted moving average --- the data moves as wells as the weights.

Sample Problem 3 (Exponential Smoothing). PC World Services assembles

customized personal computers from generic parts. The company purchases generic
computer parts in volume at a discount from a variety of sources. Thus, they need a
good forecast of demand for their computers so that they will know how many
computer component parts to purchase and stock. The purchasing officer of the
company has accumulated the demand data for the past twelve months and is
shown in the table below. The officer desires to get a forecast considering
exponential smoothing forecast with smoothing constant equal to 0.30.

PERIOD MONTH DEMAND

1 January 38
2 February 41
3 March 42
4 April 38
5 May 46
6 June 51
7 July 44
8 August 48
9 September 57
10 October 53
11 November 56
12 December 55

Page | 23
Solution

Formula for Exponential Smoothing

= −1(1 − )+ ( )

Where:
= ℎ
−1 =
= ℎ
=

To answer the above problem, start with forecasting the period 3 which is
March. Take note that there is no forecast for Period 1 (January) because
there is no data before it. The forecast for Period 2 (February) will just be
38. Here, it applies the Naïve Approach – meaning, the previous actual
value is the value of the forecast for the next period. Again, to summarize:

 No forecast for Period 1 (January)

 Use Naïve Approach to forecast Period 2 (February) which is 38.
 To forecast Period 3 (March) and succeeding periods, use Exponential
Smoothing Formula
Applying the Exponential Smoothing formula to forecast Period 3 (March), 38 will be used as the value for −1.
3 = 38(1 − 0.30) + 41(0.30)
3 = 38.90

4 = 38.90(1 − 0.30) + 42(0.30)

4 = 39.83

Just follow the pattern of computing until you reach the forecast for January
which is 52.79 or 53 computers.

Note: For us to arrive with the same answer, let’s agree to round off every
forecast value to 2 decimal places.

Page | 24
Sample Problem 4 (Linear Trend). In reference to problem no. 23, the demand
data for computers appears to follow an increasing linear trend. The company
wants to compute a linear trend line to see if it is more accurate than exponential
smoothing and adjusted exponential smoothing forecast. What is the forecast for
January?

Solution
Formula for Linear Trend

=
+

∑ −
=
2 2

− ̅

= ̅− ̅

Complete the table.

PERIOD (X) DEMAND (Y) XY X2

1 38 38 1
2 41 82 4
3 42 126 9
4 38 152 16
5 46 230 25
6 51 306 36
7 44 308 49
8 48 384 64
9 57 513 81
10 53 530 100
11 56 616 121
12 55 660 144
Sum 78 569 3945 650

78 569
̅= = 6.5 ̅=
12 12

̅= 47.42

Get first the value of b

∑ − ̅=∑2− 2
̅

Page | 25
 = 3945 − [(12)(6.5)(47.42)
650 − [12(6.5)2]
= 1.72

Get the value of a

= ̅− ̅
= 47.42 − (1.72)(6.5)
= 36.2

Find the forecast for period 13

= +
13 = 36.2 + (1.72)(13)
13 = 58.62 59 13

E. FORECAST ACCURACY

Mean Absolute Deviation (MAD)

One of the metrics used to measure the forecast accuracy is the Mean Absolute
Deviation or MAD.

How to compute MAD?

1. Get the “Deviation”. The deviation means difference. Meaning, we have to

get the individual difference between the forecast value and actual data.
This is also called Forecast Error.

2. Make the difference “Absolute”. Absolute means that the value should
always be positive (+).

3. Get the “Mean” or average of those absolute values.

Mean Absolute Percent Deviation (MAPD)

Another metric used to measure the forecast accuracy is the Mean Absolute
Percent Deviation or MAPD.

How to compute MAPD?

1. Get the “Deviation”. The deviation means difference. Meaning, we have to

get the individual difference between the forecast value and actual data.
This is also called Forecast Error.

Page | 26
 2. Make the difference “Absolute”. Absolute means that the value should
always be positive (+).

3. Get the summation of those absolute values.

4. Divide the summation of absolute values by the summation of all actual

data the multiply by 100 to make it percentage.

Important Notes: The answer to the question which method is more or most
accurate among forecasting methods depends on their respective MAD or
MAPD. The lower the MAD or MAPD, the higher the accuracy. It can only be
measured when there is existing historical data.

IV. Activity

Directions:
1. Solve the following problem(s)
2. Show complete solution.
3. Box the final answer.
4. Use separate sheet.
5. Answer should be handwritten.
6. Where and when to submit: To Be Announced (TBA).

Problems:

1. In reference to Sample Problem No. 1 (Simple Moving Average), what would be

the forecast for December?

2. In reference to Sample Problem No. 2 (Weighted Moving Average), what would

be the forecast for December?

3. In reference to Sample Problem No. 3 (Exponential Smoothing), complete the

solutions arriving at the given answer which is 52.79 or 53 computers.

4. In reference to Sample Problem No. 3 (Exponential Smoothing), find its MAD &
MAPD.

V. Assessmen
t Directions:
1. Solve the following problem(s)
2. Show complete solution.
3. Box the final answer.
4. Use separate sheet.
5. Answer should be handwritten.
6. Where and when to submit: To Be Announced (TBA).
Page | 27
 Problem

MONTH ORDERS
January 120
February 90
March 100
April 75
May 110
June 50
July 75
August 130
September 110
October 90

1. Given the historical data above, compute the forecast for the next month using:
a) 3-month SMA
b) 3-month WMA
3-2-1 will be used weights. The 3 will be assigned to the most recent data,
followed 2, then 1.

2. Determine which 2 methods is more accurate using MAD. Justify your answer.

VI. Reflection

Direction: Reflect on the following questions.

1. In your own words, what is the importance of forecasting in business operations?

VII. Assignment

1. Look for a case study involving qualitative forecasting methods. Digest the
case study, write it and just highlight the following:
a. The company’s purpose of doing the forecasting.
b. The actual effect after the company executed the forecasting.

VIII. References

 Heizer, Render (2011). Operations Management, 10th Edition.pdf, n.d.

Page | 28