Sei sulla pagina 1di 37

DECISION ANALYSIS (DA)

USING SMART
(Simple Multi Attribute Ratting Technique)
Introduction
• Decision problem that involves a number of objectives tends
to overlook making tradeoff between these objectives. It
causes these problems:
 Selection of option that perform well only for particular
objective(s)
 Rejection of attractive option since their good
performance in some objectives can’t compensate bad
performance in another objective.
Introduction
For example:
– A laptop is rejected because of its slow processor speed,
despite its compactness, reliability, and low price.
– A supplier is chosen because of his low price, despite his
slow delivery times and poor after-sales service.
Introduction
• Those problems rise since decision maker has limited
information processing capacity.
• When face a large and complex problem:
– too much information to handle simultaneously.
a decision maker is forced to make simplification

• Main role of our analysis is to enable a decision maker to gain


an increased understanding about faced problem.
– May enable a manager to reduce a large number of
alternatives.
Objectives & Attributes/Criteria
• Objectives
– Preferred direction of movement
– Use terms “minimize” or “maximize”
• Minimize cost or maximize market share.

• Attributes/Criteria
– Measure performance in relation to an objective
– If the objective “maximize the exposure of a television
advertisement”
• The attribute: “number of people surveyed who recall
seeing the advertisement”
An Office Location Problem
• A small printing & photocopying business must move to new
place.
• The owner considers 7 possible locations
Location Annual Rent ($)
A Addison Square 30,000
B Bilton Village 15,000
C Carlisle Walk 5,000
D Denver Street 12,000
E Elton Street 30,000
F Filton Village 15,000
G Gorton Square 10,000
An Office Location Problem
• Consideration toward some factors:
– Addison Square
• a prestigious location, close to potential customers
• expensive to rent
• an old dark building that won’t be comfortable for staff
to work in.
– Bilton Village
• new building that will provide excellent working
condition.
• several miles from center town, where most of
potential customers are to be found
8 SMART Stages
Simple Multi-attribute Rating Technique
• Stage 1, identify the decision maker, in our problem is the
business owner.
• Stage 2, identify the alternative courses of action, in this
problem are 7 possible locations.
• Stage 3, identify the attributes with are relevant to the
decision problem, in our case, rent cost, quality of working
condition, size deserve to consider.
– We need set of attributes that can be assessed on a
numerical scale.
• Stage 4, assign value to measure performance of each
alternative based on chosen attribute.
8 SMART Stages
Simple Multi-attribute Rating Technique
• Stage 5, determine a weight for each attribute, that reflects
their level of importance for the decision maker.
• Stage 6, for each alternative, take a weighted average of
values assigned to that alternatives
– Represents how well an office performs over all the
attributes.
• Stage 7, make a provisional decision (elimination of inferior
alternatives).
• Stage 8, perform sensitivity analysis, to see how robust the
decision.
SMART-Stage 3
• In our case, we have already passed both stages 1 & 2.
• We then jump to stage 3, to identify the attributes which the
decision maker considers to be relevant with the problem.
– Creating a value tree can be useful
criteria

costs benefits

turnover working condition

rent electricity cleaning closeness to visibility image size comfort car


customer parking
SMART-Stage 3

• Starting of categorization from both costs & benefits is an


optional.
benefits

costs benefits

benefits

short term costs long term cost convenience on the move benefits

benefits

effectiveness of the system implementation


SMART-Stage 3

• Annual rent, electricity cost, & cost of having regularly


cleaned is clearly can be assessed with numerical number.

costs

rent electricity cleaning


SMART-Stage 3
• Potential for improving turnover & staff working condition still
need to specify
• They couldn’t be understood easily
• The others may have different perception toward them
Must be more detailed

benefits

turnover working condition

closeness to visibility image size comfort car


customer parking
SMART-Stage 3
• Closeness of the office to potential customers is easily understood.
• Visibility of the site, much business is generated from people who
see the office while passing by.
• The image of the location, a decaying building in a back street may
convey a poor image and lead to a loss of business.
• Attributes size, comfort, & car parking facilities can be understood
well.
benefits

turnover working condition

closeness to visibility image size comfort car parking


customer
SMART-Stage 3

• Special attention should be given to:


– Completeness, all attributes which are relevant have been
included
– Operationality, can be evaluated & compared
– Decomposability, can be specified to more operational
level
– Absence of redundancy, no attribute duplication
– Minimum size, the smaller the size then the easier to
analyze
• it takes tradeoff between completeness & minimum
size
SMART-Stage 4
Measuring how well the options perform on each attribute
• Find out how well those locations’ performance based on
each attributes.
Location Annual Rent ($) Annual Cleaning Annual Electricity
Cost ($) Cost ($)
Addison Square 30,000 3,000 2,000

Bilton Village 15,000 2,000 800

Carlisle Walk 5,000 1,000 700

Denver Street 12,000 1,000 1,100

Elton Street 30,000 2,500 2,300

Filton Village 15,000 1,000 2,600

Gorton Square 10,000 1,100 900


SMART-Stage 4
Measuring how well the options perform on each attribute
• Cost can be represented through currency used.
• Size can be measured with floor area in square feet,

• How could we measure attribute such as “image” & “comfort”?


Direct Rating
• In next Direct Rating example we will consider about measurement
of “image”.

• Value Function is used to transfer “actual value” to the order that is


more appropriate to decision maker’s preference
• Recall the case between pedicab driver & billionaire
SMART-Stage 4
Measuring how well the options perform on each attribute
1. Direct Rating (for Abstract Value)
100 Addison Square
I. Addison Square 90 Elton Street
II. Elton Street 80
70 Filton Village
III. Filton Village
60
IV. Denver Street 50
V. Gorton Square 40
VI. Bilton Village 30 Denver Street
VII. Carlisle Walk 20 Gorton Square
10 Bilton Village
a. Rank image for those 0 Carlisle Walk
location b. Assign them value within
range 0-100
We have interval scale for
the improvement
between location
SMART-Stage 4
Measuring how well the options perform on each attribute
2. Value Function (Transferring Actual Value/Dealing with DM’s Preference)
Location Size
(Floor Area in ft2)
Elton Street 1,500

Addison Square p 1,000 The larger the area


r
Denver Street e 800  the more attractive
f
Gorton Square 700
e • Carlisle Walk & Filton
Bilton Village r 550 Village have same
r
Carlisle Walk 400 preference level.
e
Filton Village d 400
For example, for the decision maker improvement from 1000 ft2 to 1,500
ft2 isn’t as attractive as 500 ft2 to 1,000 ft2.
 we should transfer actual size to decision maker’s value
SMART-Stage 4
Measuring how well the options perform on each attribute
2. Value Function (Transferring Actual Value/Dealing with DM’s Preference)
Location Size (ft2)
100 0

Elton Street 1,500 90 0


80 0
Addison Square 1,000 v
70 0
a
Denver Street 800 l 60 0
u 50 0
Gorton Square 700 e 40 0

Bilton Village 550 30 0


20 0
Carlisle Walk 400
101500 1500
0 1500 1500 1500 1500 1500 1500 1500 1500 1500 1500

Filton Village 400 0 0


400 500 600 700 800 900 1000 1100 1200 1300 1400 1500
floor area (ft2)

After interviewing the decision maker, we get:


v(400) = 0; v(500) = 25; v(700) = 50; v(1,000) = 75; v(1,500) = 100
SMART-Stage 4
Measuring how well the options perform on each attribute

If value of all attributes is converted to range [0, 100] then:


Location Closeness Visibility Image Size Comfort Car Parking

Addison Square 100 60 100 75 0 90

Bilton Village 20 80 10 30 100 30

Carlisle Walk 80 70 0 0 10 100

Denver Street 70 50 30 55 30 90

Elton Street 40 60 90 100 60 70

Filton Village 0 0 70 0 80 0

Gorton Square 60 100 20 50 50 80


SMART-Stage 5
Determining weight of the attributes
best
I. Closeness to customer 100 0

II. Visibility 90 0
best
III. Image 80 0
best
IV. Size 70 0
V. Comfort 60 0
VI. Car-parking facilities 50 0
a. Ranking level of importance of 40 0
best
each attribute 30 0
best
 assigning the most 20 0
best
preferred with weight 10 0
100. 0 0
worst worst worst worst worst worst
b. Comparing the rest attributes with the most preferred
How much important:
improvement from the worst improvement from the worst
alternative to the best compare to alternative to the best alternative
alternative of visibility, of closeness to customer
SMART-Stage 5
Determining weight of the attributes
best
100 0 Location Original Normalized
90 0 Weights Weight
best
80 0 Closeness to 100 0.32
best customers
70 0
Visibility 80 0.26
60 0
50 0 Image 70 0.23
40 0
best Size 30 0.10
30 0
best
20 0 Comfort 20 0.06
best
10 0
Car-parking 10 0.03
0 0 facilities
worst worst worst worst worst worst
closeness visibility image size comfort car-parking
to facilities
customer

normalized weight = original weight * 100% (rounded to nearest number)


total original weight
SMART-Stage 6
Calculating aggregate of weighted value

• Total weighted value for location Addison Square


Attribute Value Normalized Value x
Weight Normalized
Weight
Closeness to customers 100 0.32 32

Visibility 60 0.26 15.6

Image 100 0.23 23

Size 75 0.10 7.5

Comfort 0 0.06 0

Car-parking facilities 90 0.03 2.7

Total 80.8
SMART-Stage 6
Calculating aggregate of weighted value

Location Closeness Visibility Image Size Comfort Car Parking Aggregate of


(0.32) (0.26) (0.23) (0.10) (0.06) (0.03) Weighted Value

Addison Square 100 60 100 75 0 90 80.8

Bilton Village 20 80 10 30 100 30 39.4

Carlisle Walk 80 70 0 0 10 100 47.4

Denver Street 70 50 30 55 30 90 52.3

Elton Street 40 60 90 100 60 70 64.8

Filton Village 0 0 70 0 80 0 20.9

Gorton Square 60 100 20 50 50 80 60.2


SMART-Stage 7
Provisional decision: trading benefit vs cost
2. If the owner has difficulty in comparing the weight for both benefit & cost

Total Aggregated
Location
Cost ($) Benefit
100 0
A Addison Square 35,000 80.8 90 0
A
80 0 x
B Bilton Village 17,800 39.4 E
70 0 x
G

value of benefit
C Carlisle Walk 6,700 47.4 60 0 x
D
x C
50 0 x
D Denver Street 14,100 52.3 B
40 0 x
E Elton Street 34,800 64.8 30 0 F
x
20 0
F Filton Village 18,600 20.9
1500 10 0
G Gorton Square 12,000 60.2 0 0
40,000 30,000 20,000 10,000 0
cost

: efficient frontier : preferred direction : dominated area


SMART-Stage 8: Sensitivity Analysis

Sensitivity analysis is:


• a method to observe stability of solutions when level of
importance for particular attributes is changed.
• a gauge to measure robustness of solutions.
• the study of effect of input perturbation toward output.
SMART-Stage 8: Sensitivity Analysis

• Sensitivity analysis between turnover vs working condition

benefits

costs benefits

turnover working condition

rent electricity cleaning closeness to visibility image size comfort car


customer parking
SMART-Stage 8: Sensitivity Analysis
Turnover = 0
Location Original Normalized
Weights Weight
Closeness to 0 0
customers
Visibility 0 0

Image 0 0

Size 30 0.5

Comfort 20 0.333

Car-parking 10 0.167
facilities

normalized weight = original weight * 100% (rounded to nearest number)


total original weight
SMART-Stage 8: Sensitivity Analysis
Location Closeness Visibility Image Size Comfort Car Parking Aggregate of
(0.32) (0.26) (0.23) (0.10) (0.06) (0.03) Weighted Value
Addison Square 100 60 100 75 0 90 80.8
Bilton Village 20 80 10 30 100 30 39.4
Carlisle Walk 80 70 0 0 10 100 47.4
Denver Street 70 50 30 55 30 90 52.3
Elton Street 40 60 90 100 60 70 64.8
Filton Village 0 0 70 0 80 0 20.9
Gorton Square 60 100 20 50 50 80 60.2

Location Closeness Visibility Image Size Comfort Car Parking Aggregate of


(0.00) (0.00) (0.00) (0.50) (0.333) (0.167) Weighted Value
Addison Square 100 60 100 75 0 90 52.53
Bilton Village 20 80 10 30 100 30 53.31
Carlisle Walk 80 70 0 0 10 100 20.03
Denver Street 70 50 30 55 30 90 52.52
Elton Street 40 60 90 100 60 70 81.67
Filton Village 0 0 70 0 80 0 26.64
Gorton Square 60 100 20 50 50 80 55.01
SMART-Stage 8: Sensitivity Analysis
Working condition = 0
Location Original Normalized
Weights Weight
Closeness to 100 0.4
customers
Visibility 80 0.32

Image 70 0.28

Size 0 0

Comfort 0 0

Car-parking 0 0
facilities

normalized weight = original weight * 100% (rounded to nearest number)


total original weight
SMART-Stage 8: Sensitivity Analysis
Location Closeness Visibility Image Size Comfort Car Parking Aggregate of
(0.32) (0.26) (0.23) (0.10) (0.06) (0.03) Weighted Value
Addison Square 100 60 100 75 0 90 80.8
Bilton Village 20 80 10 30 100 30 39.4
Carlisle Walk 80 70 0 0 10 100 47.4
Denver Street 70 50 30 55 30 90 52.3
Elton Street 40 60 90 100 60 70 64.8
Filton Village 0 0 70 0 80 0 20.9
Gorton Square 60 100 20 50 50 80 60.2

Location Closeness Visibility Image Size Comfort Car Parking Aggregate of


(0.40) (0.32) (0.28) (0.00) (0.00) (0.00) Weighted Value
Addison Square 100 60 100 75 0 90 87.2
Bilton Village 20 80 10 30 100 30 36.4
Carlisle Walk 80 70 0 0 10 100 54.4
Denver Street 70 50 30 55 30 90 52.4
Elton Street 40 60 90 100 60 70 60.4
Filton Village 0 0 70 0 80 0 19.6
Gorton Square 60 100 20 50 50 80 61.6
Location Aggregate of Weighted Value Aggregate of Weighted Value Aggregate of Weighted Value
(Turnover = 0) (Original) (Working Condition = 0)
Addison Square 52.53 80.8 87.2
Bilton Village 53.31 39.4 36.4
Carlisle Walk 20.03 47.4 54.4
Denver Street 52.52 52.3 52.4
Elton Street 81.67 64.8 60.4
Filton Village 26.64 20.9 19.6
Gorton Square 55.01 60.2 61.6

100 0 100 0
90 0 90 0
80 0 80 0
Value of benefit

Value of benefit
70 0 70 0
60 0 60 0
50 0 50 0
40 0 40 0
30 0 30 0
20 0 20 0
101500 1500
0 1500 1500 1500 1500 1500 1500 1500 1500 1500 10 0
1500
0 0 0
0 10 20 30 40 50 60 70 80 90 100
weight placed on turnover
SMART-Stage 8: Sensitivity Analysis

100 0 100 0
90 0 90 0
80 0 80 0
Value of benefit

Value of benefit
70 0 70 0
60 0 60 0
50 0 50 0
40 0 40 0
30 0 30 0
20 0 20 0
101500 1500
0 1500 1500 1500 1500 1500 1500 1500 1500 1500 10 0
1500
0 0 0
0 10 20 30 40 50 60 70 80 90 100
weight placed on turnover

• As long weight of total turnover < 52.1  Elton Street is the most attractive
• When weight of total turnover > 52.1  Addison Square is the most attractive
PPA for Addison Square
POTENTIAL CONSEQUENCE POSSIBLE PREVENTIVE CONTINGENT
PROBLEM CAUSE ACTION ACTION
Cannot pay Expelled from Rent is so Negotiate for a Rent smaller
rent property expensive longer term, space for front
lower rent office and
move back
office to
cheaper
location
Low cash flow Create Change
after moving to promotional business model
new office program
Uncomfortable High employee Old dark Refurbish the Move to
office turnover building office with another
bright color location
and new
furniture
Thanks

Potrebbero piacerti anche