topic2 forecasting

1

INDUSTRIAL MANAGEMENT (MEM 575)

Topic 2: Forecasting

Lecturer: Pn. Wan Mazlina Wan Mohamed Office: T1-A11-11A

MEM 575 2

What is Forecas3ng?

þ  Process of predic3ng a future event

þ  Underlying basis of all business decisions þ  Produc3on þ  Inventory þ  Personnel þ  Facili3es

Hmm…. you gonna get an A for this subject

MEM 575 3

þ  Short-‐range forecast þ  Up to 1 year, generally less than 3 months þ  Purchasing, job scheduling, workforce levels, job

assignments, produc3on levels þ  Medium-‐range forecast

þ  3 months to 3 years þ  Sales and produc3on planning, budge3ng

þ  Long-‐range forecast þ  3+ years þ  New product planning, facility loca3on, research

and development

Forecas3ng Time Horizons

MEM 575 4

Seven Steps in Forecas3ng þ Determine the use of the forecast þ  Select the items to be forecasted þ Determine the 3me horizon of the forecast þ  Select the forecas3ng model(s) þ Gather the data þ Make the forecast þ Validate and implement results

MEM 575 5

Types of Forecasts

þ  Economic forecasts þ  Address business cycle – infla3on rate, money

supply, housing starts, etc.

þ  Technological forecasts þ  Predict rate of technological progress þ  Impacts development of new products

þ  Demand forecasts þ  Predict sales of exis3ng products and services

MEM 575 6

Strategic Importance of Forecas3ng

þ  Human Resources – Hiring, training, laying off workers

þ  Capacity – Capacity shortages can result in undependable delivery, loss of customers, loss of market share

þ  Supply Chain Management – Good supplier rela3ons and price advantages

MEM 575 7

The Reali3es!

þ Forecasts are seldom perfect þ Most techniques assume an

underlying stability in the system þ Product family and aggregated

forecasts are more accurate than individual product forecasts

MEM 575 8

Forecas3ng Approaches

þ Used when situa3on is vague and liale data exist þ  New products þ  New technology

þ  Involves intui3on, experience þ  e.g., forecas3ng sales on Internet

Qualita3ve Methods

MEM 575 9

Forecas3ng Approaches

þ Used when situa3on is ‘stable’ and historical data exist þ  Exis3ng products þ  Current technology

þ  Involves mathema3cal techniques þ  e.g., forecas3ng sales of color televisions

Quan3ta3ve Methods

MEM 575 10

Overview of Quan3ta3ve Approaches

1.  Naive approach 2.  Moving averages 3.  Exponen3al smoothing 4.  Trend projec3on 5.  Linear regression

Time-‐Series Models

Associa3ve Model

MEM 575 11

þ  Set of evenly spaced numerical data þ  Obtained by observing response variable at

regular 3me periods

þ  Forecast based only on past values, no other variables important þ  Assumes that factors influencing past and

present will con3nue influence in future

Time Series Forecas3ng

MEM 575 12

Components of Demand De

mand for p

rodu

ct or service

| | | | 1 2 3 4 Year

Average demand over four years

Seasonal peaks

Trend component

Actual demand

Random varia3on

Figure 4.1

MEM 575 13

þ  Persistent, overall upward or downward paaern

þ Changes due to popula3on, technology, age, culture, etc.

þ  Typically several years dura3on

Trend Component

NY -‐ KJP 585 2009 14

þ  Regular paaern of up and down fluctua3ons

þ  Due to weather, customs, etc. þ  Occurs within a single year

Seasonal Component

Number of Period Length Seasons

Week Day 7 Month Week 4-‐4.5 Month Day 28-‐31 Year Quarter 4 Year Month 12 Year Week 52

MEM 575 15

þ Repea3ng up and down movements þ Affected by business cycle, poli3cal, and

economic factors þ Mul3ple years dura3on þ Oken causal or

associa3ve rela3onships

Cyclical Component

0 5 10 15 20

MEM 575 16

þ  Erra3c, unsystema3c, ‘residual’ fluctua3ons

þ  Due to random varia3on or unforeseen events

þ  Short dura3on and nonrepea3ng

Random Component

M T W T F

MEM 575 17

Naive Approach

þ Assumes demand in next period is the same as demand in most recent period þ e.g., If January sales were 68, then

February sales will be 68 þ  Some3mes cost effec3ve and efficient þ Can be good star3ng point

MEM 575 18

þ  Moving average þ Weighted moving average þ  Exponen3al smoothing

Techniques for Averaging

MEM 575 19

Moving Average Method

Moving average = ∑ demand in previous n periods

n

A forecasting method that uses an average of the ‘n’ most recent periods of data to forecast the next period. Useful if we can assume that market demands will stay fairly steady over time.

e.g. a 4-month moving average is found by summing the demand during the past 4 months and dividing by 4. This practice tends to smooth out short term irregularities in the data series.

The above is used as an estimate of the next period’s demand

MEM 575 20

Moving Average Example

Storage shed sales at a Garden Supply shop are as shown in the following Table.

Example 1:

Calculate the 3-month moving average forecast.

MEM 575 21

January 10 February 12 March 13 April 16 May 19 June 23 July 26

Actual 3-Month Month Shed Sales Moving Average

(12 + 13 + 16)/3 = 13 2/3 (13 + 16 + 19)/3 = 16 (16 + 19 + 23)/3 = 19 1/3


10 12 13

(10 + 12 + 13)/3 = 11 2/3

MEM 575 22


e.g. the forecast for December is 20.7 The forecast for coming January is (18+16+14)/3=16.0

MEM 575 23

Graph of Moving Average

| | | | | | | | | | | | J F M A M J J A S O N D

Shed

Sal

es

30 – 28 – 26 – 24 – 22 – 20 – 18 – 16 – 14 – 12 – 10 –

Actual Sales

Moving Average Forecast

MEM 575 24

Weighted Moving Average

Weighted moving average =

∑ (weight for period n) x (demand in period n)

∑ weights

When a detectable trend or pattern is present, weights can be used to place more emphasis on recent values. This makes forecasting techniques more responsive to changes because more recent periods may be more heavily weighted.

MEM 575 25

Weighted Moving Average Ex: Example 2 The shop in Example 1 decides to forecast storage shed sales by weigh3ng the past 3 months as

follows: Period Weight applied Last month 3 2 months ago 2 3 months ago 1 _____________________________ Solu3on: ∑ (weights) = 6 Based on the weigh3ngs above, the forecast for any month

[(3 x Sales last month) + (2 x Sales 2 months ago) + (1 x Sales 3 months ago)] = -‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐

∑ (weights)

MEM 575 26

January 10 February 12 March 13 April 16 May 19 June 23 July 26

Actual 3-Month Weighted Month Shed Sales Moving Average

[(3 x 16) + (2 x 13) + (12)]/6 = 141/3 [(3 x 19) + (2 x 16) + (13)]/6 = 17 [(3 x 23) + (2 x 19) + (16)]/6 = 201/2

10 12 13

[(3 x 13) + (2 x 12) + (10)]/6 = 121/6

Weights Applied Period 3 Last month 2 Two months ago 1 Three months ago 6 Sum of weights

MEM 575 27

Weighted Moving Average Ex:

Note that in this situa3on more heavily weigh3ng the latest month provides a much more accurate projec3on. Note also that moving averages are effec3ve in smoothing out sudden fluctua3ons in the demand paaern to provide stable es3mates.

MEM 575 28

Moving Average And Weighted Moving Average

Note from the graph that both moving averages lag the actual demand. The weighted moving average, however reacts more quickly to changes in demand.

MEM 575 29

þ  Increasing n smooths the forecast but makes it less sensi3ve to changes

þ Do not forecast trends well þ Require extensive historical data

Poten3al Problems With Moving Average

MEM 575 30

Exponen3al Smoothing

• Is a weighted moving average forecasting technique in which data points are weighted by an exponential function. • This technique involves little record keeping of past data.

MEM 575 31

Exponen3al Smoothing

New forecast = Last period’s forecast + α (Last period’s actual demand – Last period’s forecast)

Ft = Ft – 1 + α(At – 1 - Ft – 1)

where Ft = new forecast Ft – 1 = previous forecast

α = smoothing (or weighting) constant (0 ≤ α ≤ 1)

Remember This!!!!!!!! Basic exponential smoothing formula:

MEM 575 32

Exponen3al Smoothing Example

Example 3 In January, a car dealer predicted February demand for 142 Ford Mustangs. Actual February demand was 153. Using a smoothing constant chosen by management of α = 0.20, forecast the March demand using the exponential smoothing model. Solution: Substituting into the formula above, New forecast (for March demand), FMac = FFeb + α (AFeb – FFeb)

= 142 + 0.20 (153 – 142) = 144.2

Therefore the March demand forecast for Ford Mustang is 144.

MEM 575 33

Measuring Forecast Error Forecast error (or Deviation) = Actual demand – Forecast demand = At - Ft. Several measures in use: • Mean absolute deviation (MAD) • Mean squared error (MSE) • Mean absolute percent error (MAPE)

∑ | Actual - Forecast | MAD = ------------------------------

n ∑ (Forecast error)2

MSE = ------------------------ n n 100 ∑ | Actual i - Forecast i | / Actual i

MAPE = -----------i=1-------------------------------------------- n

MEM 575 34

Trend Projec3on

•  A time-series forecasting method that fits a trend line to a series of historical data points and then projects the line into the future for forecasts.

•  It is usually for medium-to-long range forecasts.

MEM 575 35

Forecast Error Example:

MEM 575 36

MEM 575 37

MEM 575 38

MEM 575 39

Exponen3al Smoothing Example 2

Demand for the last four months was:

Predict demand for July using each of these methods: (A) 1) A 3-period moving average 2) exponential smoothing with alpha equal to .20 (use naïve to

begin). (B) 3) If the naive approach had been used to predict demand for April

through June, what would MAD have been for those months?

MEM 575 40

Exponen3al Smoothing Example 2

Month Demand Forecast

March 6 -‐

April 8 6

May 10 6 + 0.2(8 – 6) = 6.4

June 8 6.4 + 0.2(10 – 6.4) = 7.12

7.12 + 0.2(8 – 7.12) = 7.296

A)  1. (8+10+8)/3 = 8.33 (July Forecast) 2. Use naïve to begin

B) Month March April May June

Demand 6 8 10 8

Naïve -‐ 6 8 10

Error -‐ +2 +2 -‐2

MAD 6/3 = 2.0

MEM 575 41

Moving Average Weekly sales of ten-‐grain bread at the local organic food market are in the table below. Based on this data, forecast week 9 using a five-‐week moving average.

Other Examples

Week 1 2 3 4 5 6 7 8

Sales 415 389 420 382 410 432 405 421

MEM 575 42

Exponential Smoothing & MAD Jim's department at a local department store has tracked the sales of a product over the last ten weeks. Forecast demand using exponential smoothing with an alpha of 0.4, and an initial forecast of 28.0. Calculate MAD.

Other Examples

Period Demand 1 24 2 23 3 26 4 36 5 26 6 30 7 32 8 26 9 25

10 28

MEM 575 43

Period Demand Forecast Error Absolute 1 24 28.00 2 23 26.40 -3.40 3.40 3 26 25.04 0.96 0.96 4 36 25.42 10.58 10.58 5 26 29.65 -3.65 3.65 6 30 28.19 1.81 1.81 7 32 28.92 3.08 3.08 8 26 30.15 -4.15 4.15 9 25 28.49 -3.49 3.49

10 28 27.09 0.91 0.91 Total 2.64 32.03

Average 0.29 3.56 Bias MAD

Other Examples – Exponen3al Smoothing

topic2 forecasting

Documents