Sei sulla pagina 1di 53

MEE1014

INDUSTRIAL ENGINEERING AND MANAGEMENT


Module I

Introduction to macro and micro economics:


Macro economic measures – micro economics – Demand and supply –
Determinants of demand and supply – Elasticity of demand – Demand
forecasting techniques (short term & long term) – Problems.

Dr K Jayakrishna, Associate Professor, SMEC-VIT


Forecasting

Forecasting is the use of historic data to determine the direction of future trends.

Businesses utilize forecasting to determine how to allocate their budgets or plan for
anticipated expenses for an upcoming period of time.

This is typically based on the projected demand for the goods and services they offer.

Dr K Jayakrishna, Associate Professor, SMEC-VIT


Characteristics of Forecasts

1. Forecasts are normally wrong since it is an approximation of demand


estimates in the future.
2. A good forecast has more than a single number, as it includes
 A mean value and standard deviation
 An accuracy range (high and low)
3. Aggregated forecasts are usually more accurate because it can adjust the
variation in actual demand of individual products.
4. Accuracy erodes as we go further into the future because the increase in time
horizon increases the uncertainty in demand pattern.
5. Forecasts should not be used to the exclusion of known information.
Dr K Jayakrishna, Associate Professor, SMEC-VIT
Forecasting Horizons
 Short-range forecast Quantitative
Usually < 3 months methods
Detailed use of
Job scheduling, worker assignments system
 Medium-range forecast
3 months to 2 years
Sales/production planning
 Long-range forecast
Design of
> 2 years Qualitative
system
New product planning Methods

Dr K Jayakrishna, Associate Professor, SMEC-VIT


Steps of Forecasting

1. Determination of objectives of forecasting


2. Selection of the items to be forecast
3. Determination of the time horizon of the forecast
4. Selection of the suitable forecasting model
5. Collection of the data
6. Validation of the forecasting model
7. Forecasting the demand
8. Implementation of the results.

Dr K Jayakrishna, Associate Professor, SMEC-VIT


Forecasting methods

Forecasting methods are divided into two categories:


 Qualitative forecasting
 Quantitative forecasting

Dr K Jayakrishna, Associate Professor, SMEC-VIT


Qualitative Forecasting

• Qualitative forecasting is an estimation methodology that uses expert judgment,


rather than numerical analysis.

• This type of forecasting relies upon the knowledge of highly experienced


employees and consultants to provide insights into future outcomes.

• This approach is substantially different from quantitative forecasting, where


historical data is compiled and analysed to discern future trends.

Dr K Jayakrishna, Associate Professor, SMEC-VIT


Qualitative Forecasting

• Qualitative forecasting is most useful in situations where it is suspected that


future results will depart markedly from results in prior periods, and which
therefore cannot be predicted by quantitative means.

• For example, the historical trend in sales may indicate that sales will increase
again in the next year, which would normally be measured using trend line
analysis; however, an industry expert points out that there will be a materials
shortage at a key supplier that will force sales downward.

Dr K Jayakrishna, Associate Professor, SMEC-VIT


Qualitative Forecasting

• Another situation in which qualitative forecasting can be useful is in the


assimilation of large amounts of narrowly-focused local data to discern trends
that a more quantitative analysis might not find.

• For example, a construction company needs to know what style of home to build
in a certain area, and relies on a local population expert to find out that the area in
question is being abandoned by younger families and replaced by an older,
retirement-age group. Consequently, the builder constructs smaller one-level
homes with fewer bedrooms.

Dr K Jayakrishna, Associate Professor, SMEC-VIT


Qualitative Forecasting

• This approach also works well when a course of action must be derived from
inadequate data. In this case, a qualitative analysis will seek to link disparate data
to construct a more broad-based view, sometimes incorporating intuition to
construct this view.

Dr K Jayakrishna, Associate Professor, SMEC-VIT


Qualitative Forecasting

Executive committee
consensus

Qualitative Forecasting
Delphi method

Survey of sales force

Expert evaluation
technique

Dr K Jayakrishna, Associate Professor, SMEC-VIT


Qualitative Forecasting

Executive Judgment: Opinion of a group of high level experts or managers is pooled

Sales Force Composite: Each regional salesperson provides his/her sales estimates.
Those forecasts are then reviewed to make sure they are realistic. All regional forecasts
are then pooled at the district and national levels to obtain an overall forecast.

Market Research/Survey: Solicits input from customers pertaining to their future


purchasing plans. It involves the use of questionnaires, consumer panels and tests of new
products and services.
Dr K Jayakrishna, Associate Professor, SMEC-VIT
Qualitative Forecasting
Delphi Method: As opposed to regular panels where the individuals involved are in
direct communication, this method eliminates the effects of group potential dominance
of the most vocal members. The group involves individuals from inside as well as
outside the organization.

Typically, the procedure consists of the following steps:


• Each expert in the group makes his/her own forecasts in form of
statements
• The coordinator collects all group statements and summarizes them
• The coordinator provides this summary and gives another set of questions
to each group member including feedback as to the input of other experts.
• The above steps are repeated until a consensus is reached.
Dr K Jayakrishna, Associate Professor, SMEC-VIT
Advantages of qualitative forecasting

• Ability to predict changes in sales patterns


• Allow decision makers to incorporate rich data sources consisting of their intuition,
experience and expert judgement.

Dr K Jayakrishna, Associate Professor, SMEC-VIT


Disadvantages of qualitative forecasting

• Ability to forecast accurately suffers due to;


• Considers only readily available and/or recently perceived information
• Unable to process large amounts of complex information
• Overconfident about their ability to forecast
• Political factors
• Infer relationships or patterns in data
• Influenced by other previous forecasts

Dr K Jayakrishna, Associate Professor, SMEC-VIT


Quantitative Forecasting

A statistical technique for making projections about the future which uses numerical
facts and prior experience to predict upcoming events.

The two main types of quantitative forecasting used by business analysts are;
• Time series method that uses past trends to make forecasts.
• Explanatory method that attempts to correlate two or more variables and

Dr K Jayakrishna, Associate Professor, SMEC-VIT


Quantitative Forecasting
Quantitative
Forecasting

Time Series Causal


Models Models

Moving Exponential Trend Regression


Average Smoothing Models

Dr K Jayakrishna, Associate Professor, SMEC-VIT


What is a Time Series ?

Time series data is a sequence of observations.


Time series is dynamic, it does change over time
Set of evenly spaced numerical data obtained by observing response variable at
regular time periods
Forecast based only on past values, assumes that factors influencing past, present
and future will continue
Example
Year: 2013 2014 2015 2016 2017
Sales: 78.7 63.5 89.7 93.2 92.1
Dr K Jayakrishna, Associate Professor, SMEC-VIT
Time Series Components

Trend Cyclical

Seasonal Irregular
Dr K Jayakrishna, Associate Professor, SMEC-VIT
Trend Component

Persistent, overall upward or downward pattern


Due to population, technology etc.
Data taken over a period of years

Dr K Jayakrishna, Associate Professor, SMEC-VIT


Cyclical Component

Repeating up and down movements/swings


Due to interactions of factors influencing economy
Usually 2-10 years duration
May Vary in Length

Dr K Jayakrishna, Associate Professor, SMEC-VIT


Seasonal Component

Regular pattern of up and down fluctuations/swings


Due to weather, customs etc.
Occurs within one year

Dr K Jayakrishna, Associate Professor, SMEC-VIT


Irregular Component

Erratic, unsystematic, ‘residual’ fluctuations


Due to random variation or unforeseen events
• Union strike
• War
Short duration and nonrepeating

Dr K Jayakrishna, Associate Professor, SMEC-VIT


Time series forecasting
Time Series

NO YES
Trend
Smoothing methods Trend models
?

Moving average Exponential Smoothing

Linear Quadratic Exponential Autoregressive

Dr K Jayakrishna, Associate Professor, SMEC-VIT


Plotting time series data

Dr K Jayakrishna, Associate Professor, SMEC-VIT


Time series forecasting
Time Series

NO YES
Trend
Smoothing methods Trend models
?

Moving average Exponential Smoothing

Linear Quadratic Exponential Autoregressive

Dr K Jayakrishna, Associate Professor, SMEC-VIT


Measures of Forecast Accuracy

Dt = Demand for the period t

Ft = Forecast demand for the period t and

Then, Forecast error for the period t , (et) = ?

et = Dt - Ft
Measures of Forecast Accuracy

Mean Squared Error (MSE)

• The average of the squared forecast errors for the historical data is calculated.

• The forecasting method or parameter(s) which minimize this mean squared error
is then selected.
n

 (D t  Ft ) 2
MSE  i 1

n = Number of time period used


Measures of Forecast Accuracy

Mean Absolute Deviation (MAD)

• The mean of the absolute values of all forecast errors is calculated, and the
forecasting method or parameter(s) which minimize this measure is selected.

• The mean absolute deviation measure is less sensitive to individual large forecast
errors than the mean squared error measure.

σ𝑛𝑖=1 𝐷𝑡 − 𝐹𝑡
𝑀𝐴𝐷 =
𝑛

n = Number of time period used


Measures of Forecast Accuracy

Mean Absolute Percentage Error (MAPE)

• It is the mean of the percentage deviations of the forecast demands from the actual
demands

1 n Dt  Ft
MAPE   x100
n i 1 Dt
n = Number of time period used
Measures of Forecast Accuracy

Mean Forecast Error (MFE)

• It is the mean of the deviations of the forecast demands from the actual demands

σ𝑛𝑡=1 𝐷𝑡 − 𝐹𝑡
𝑀𝐹𝐸 =
𝑛

n = Number of years (time period used)


Naïve Method

Number of periods used for smoothing is 1.

The forecast for a period t, is simply the observed value of the previous period, t-1.
It can only be used to forecast up to one period in the future
Used only for its simplicity

Dr K Jayakrishna, Associate Professor, SMEC-VIT


Simple Moving Average Method (SMAM)
1
𝐹𝑡+1 = 𝑛
𝐷𝑡 + 𝐷𝑡−1 + ⋯ + 𝐷𝑡+1−𝑛

𝑡
1
𝐹𝑡+1 = ෍ 𝐷𝑖 𝐷
𝑛
𝑖=𝑡+1−𝑛

t = Current period

t+1 = Forecast period

F = Forecast

Dt = Actual demand in period t

n= Number of the averaging period

Dr K Jayakrishna, Associate Professor, SMEC-VIT


Simple Moving Average Method (SMAM)

The monthly demands for an office furniture (in units) are given in Table below.
Forecast the demand using 3-period .

Month
1 2 3 4 5 6 7 8 9 10 11 12
(t)
Demand
600 628 670 735 809 870 800 708 842 870 739 -
(Dt)

Dr K Jayakrishna, Associate Professor, SMEC-VIT


Simple Moving Average Method (SMAM)
1
𝐹𝑡+1 = 𝑛
𝐷𝑡 + 𝐷𝑡−1 + ⋯ + 𝐷𝑡+1−𝑛

𝑡
1
𝐹𝑡+1 = ෍ 𝐷𝑖 𝐷
𝑛
𝑖=𝑡+1−𝑛

t = Current period

t+1 = Forecast period

F = Forecast

Dt = Actual demand in period t

n= Number of the averaging period

Dr K Jayakrishna, Associate Professor, SMEC-VIT


Simple Moving Average Method (SMAM)
Time Demand for month Moving Average
(month (t)) (Dt) M(t)
1 600
2 628
3 670
4 735 633 Three period moving average t = 3
5 809 678 𝐹𝑡+1
6 870 738 1
= 𝐷 + 𝐷𝑡−1 + ⋯ + 𝐷𝑡+1−𝑛
7 800 805 𝑛 𝑡
8 708 826 1
9 842 793 𝐹3+1 = 600 + 628 + 670
3
10 870 783
11 739 807 𝐹4 = 632.66 ≈ 633
12 - 817

Dr K Jayakrishna, Associate Professor, SMEC-VIT


Double Moving Average Method (DMAM)

Just repeat simple moving average method again….

Dr K Jayakrishna, Associate Professor, SMEC-VIT


Double Moving Average Method (DMAM)

The monthly demands for an office furniture (in units) are given in Table below.
Forecast the demand for 3-periods using double moving average method.

Month
1 2 3 4 5 6 7 8 9 10 11 12
(t)
Demand
600 628 670 735 809 870 800 708 842 870 739 -
(Dt)

Dr K Jayakrishna, Associate Professor, SMEC-VIT


Double Moving Average Method (SMAM)
Double Moving
Time Demand for Moving Average
Average M(t)
(month (t)) month (Dt) M(t)

1 600 Three period moving average t = 3


𝐹𝑡+1
2 628 1
3 670 = 𝐷 + 𝐷𝑡−1 + ⋯ + 𝐷𝑡+1−𝑛
𝑛 𝑡
4 735 633
5 809 678 1
𝐹6+1 = 633 + 678 + 738
6 870 738 3
7 800 805 683 𝐹7 = 683
8 708 826 741
9 842 793 790
10 870 783 808
11 739 807 801
12 - 817 795
Dr K Jayakrishna, Associate Professor, SMEC-VIT
Weighted Moving Average Method (WMAM)
𝐹𝑡+1 = 𝑊1 𝐷𝑡 + 𝑊2 𝐷𝑡−1 … … . . +𝑊𝑛 𝐷𝑡+1−𝑛

𝑊1 > 𝑊2 >. . . . > 𝑊𝑛


t = Current period

t+1 = Forecast period

F = Forecast

Wt = Weight applied to period t

n= Total number of periods

Largest weight = More recent period…………..Lowest weight = Last period


Dr K Jayakrishna, Associate Professor, SMEC-VIT
Weighted Moving Average Method (WMAM)

The monthly demands for an office furniture (in units) are given in Table below.
Forecast the demand using 3-period WMAM.
The most recent data should be given 50 percent weightage, second year past data,
30 percent, and third year past data, 20 percent.

Month
1 2 3 4 5 6 7 8 9 10 11 12
(t)
Demand
600 628 670 735 809 870 800 708 842 870 739 -
(Dt)

Dr K Jayakrishna, Associate Professor, SMEC-VIT


Weighted Moving Average Method (WMAM)
𝐹𝑡+1 = 𝑊1 𝐷𝑡 + 𝑊2 𝐷𝑡−1 … … . . +𝑊𝑛 𝐷𝑡+1−𝑛

𝑊1 > 𝑊2 >. . . . > 𝑊𝑛

𝑊1 = 0.5 > 𝑊2 = 0.3 > 𝑊3 = 0.2


t = Current period
t+1 = Forecast period
F = Forecast
Wt = Weight applied to period t
n= Total number of periods = 3
Largest weight = More recent period…………..Lowest weight = Last period

Dr K Jayakrishna, Associate Professor, SMEC-VIT


Weighted Moving Average Method (WMAM)
Time in Weighted Moving Three period moving average t = 3
Deamnd (Dt)
months (t) Average M(t) 𝐹𝑡+1
1 600 = 𝑊1 𝐷𝑡
2 628 + 𝑊2 𝐷𝑡−1 … … . . +𝑊𝑛 𝐷𝑡+1−𝑛
3 670
𝑊1 > 𝑊2 >. . . . > 𝑊𝑛
4 735 643
5 809 694 𝑊1 = 0.5 > 𝑊2 = 0.3 > 𝑊3 = 0.2
6 870 759
7 800 825 𝐹3+1 = 0.5 × 670 + 0.3 × 628 +
8 708 823 0.2 × 600
9 842 768
𝐹4 = 643.4 ≈ 644
10 870 793
11 739 829
12 799

Dr K Jayakrishna, Associate Professor, SMEC-VIT


Time series forecasting
Time Series

NO YES
Trend
Smoothing methods Trend models
?

Moving average Exponential Smoothing

Linear Quadratic Exponential Autoregressive

Dr K Jayakrishna, Associate Professor, SMEC-VIT


Exponential Smoothing Method (ESM)
𝐹𝑡 = 𝐹𝑡−1 + 𝛼(𝐷𝑡−1 − 𝐹𝑡−1 )

Where

Ft = Smoothed average forecast for period t

Ft-1= Previous period forecast

α = Smoothing constant, weight given to previous data (0≤α≤1)

(Preferred range 0.1 to 0.3)

Dt-1=previous period demand

Dr K Jayakrishna, Associate Professor, SMEC-VIT


Exponential Smoothing Method (ESM)

The monthly demands for an office furniture (in units) are given in Table below.
Forecast the demand for the periods using the exponential smoothing method ,
considering the Smoothing constant as, α = 0.3

Month
1 2 3 4 5 6 7 8 9 10 11 12
(t)
Demand
600 628 670 735 809 870 800 708 842 870 739 -
(Dt)

Dr K Jayakrishna, Associate Professor, SMEC-VIT


Exponential Smoothing Method (ESM)
Time in
Deamnd (Dt) Forecast, Ft
months (t)
𝐹𝑡 = 𝐹𝑡−1 + 𝛼(𝐷𝑡−1 − 𝐹𝑡−1 )
1 600
2 628 600 ∝ = 0.3
3 670 608
𝐹2 = 600 + 0.3 600 − 600 = 600
4 735 627
5 809 659 𝐹3 = 600 + 0.3 628 − 600 = 608.4
6 870 704
7 800 754 𝐹4 = 608.4 + 0.3 670 − 608.4 = 626.8
8 708 768
9 842 750
10 870 777
11 739 805
12 785
Dr K Jayakrishna, Associate Professor, SMEC-VIT
Simple Linear Regression Method (SLRM)
Linear regression is based on

• Fitting a straight line to data


• Explaining the change in one variable through changes in other variables.
• In simple regression, only one independent variable is used, where as in multiple
regression two or more independent variables are involve
• By using linear regression, we are trying to explore which independent variables
affect the dependent variable

dependent variable d= a + b  (independent variable)


a=intercept, b=slope

Dr K Jayakrishna, Associate Professor, SMEC-VIT


Simple Linear Regression Method (SLRM)
• Identify dependent (y) and independent (x) variables

• Solve for the slope of the line

σ 𝑥𝑦 − 𝑛 𝑦ത 𝑥ҧ
𝑏=
σ 𝑥 2 − 𝑛 (𝑥)ҧ 2

• Solve for the y intercept


𝑎 = 𝑦ത − 𝑏𝑥ҧ

• Develop your equation for the trend line

𝑦 = 𝑎 + 𝑏. 𝑥
Dr K Jayakrishna, Associate Professor, SMEC-VIT
Simple Linear Regression Method (SLRM)
Where,

a – a constant
b – a coefficient of variable x
𝑥ҧ – mean value of x
𝑦ത – mean value of y
x – is the time
y – demand
n – time period

Dr K Jayakrishna, Associate Professor, SMEC-VIT


Simple Linear Regression Method (SLRM)

The weekly demands of a motorcycle by a retailer are shown in Table. Find an


equation of the regression line and estimate the demand for the 14th week.

Week (x) 1 2 3 4 5 6 7 8 9 10 11 12
Demand
420 450 460 420 500 550 480 520 610 570 600 590
(y)

Dr K Jayakrishna, Associate Professor, SMEC-VIT


Simple Linear Regression Method (SLRM)

Week (x) Demand (y) xy x2 σ 𝑥𝑦 − 𝑛 𝑦ത 𝑥ҧ


𝑏=
σ 𝑥 2 − 𝑛 (𝑥)ҧ 2
1 420 420 1
2 450 900 4 6170 78
42570 − 12 × ( )×( )
3 460 1380 9 𝑏= 12 12
650 − 12 × (78/12) 2
4 420 1680 16
5 500 2500 25 𝑏 = 17.23
6 550 3300 36
7 480 3360 49 𝑎 = 𝑦ത − 𝑏𝑥ҧ
8 520 4160 64
𝑎 = 514.1667 − 17.23 × 6.5 = 402.17
9 610 5490 81
10 570 5700 100 Therefore, the regression equation is
11 600 6600 121
12 590 7080 144 𝑦 = 402.17 + 17.23𝑥
∑x = 78 ∑y = 6170 ∑xy = 42570 ∑x = 650
2
𝑦14 = 402.17 + 17.23 × 14 = 643.39
Dr K Jayakrishna, Associate Professor, SMEC-VIT

Potrebbero piacerti anche