CLASS #3
Chapter 3- Decisions involving
multiple objectives: SMART
Decision Analysis for Management Judgment
Goodwin e Wright (2005)
Multiple Criteria approach
In general, the multiple criteria approach is inserted in the type of NOT
STRUCTURED problems, with some of the following characteristics:
- The criteria are at least two and conflicting;
- Criteria and solution alternatives are not clearly defined and the consequences
of choosing a given alternative with respect to at least one criterion are not
clearly understood;
- The criteria and the alternatives can be interconnected, in such a way that one
criterion seems to partially reflect another criterion, while the effectiveness of
the choice of an alternative depends on whether another alternative has been
chosen or not, in which the alternatives are not mutually exclusive;
- Solution for the problems depends on a group of people, with different points
of view and culture, often with conflicting ideas;
- The restrictions of the problem are not well defined, and there may even be
some doubt about what is a criterion and what is a restriction;
- Some criteria are quantifiable while others are based on value judgments
made on a scale;
Decisions involving multiple
objectives
This chapter will explore how decision analysis can
be used to support decision makers who have
multiple objectives. The idea is to split the problem
into small parts focusing on each part separately
The method explained here is normally applied in
situation where a particular course of action is
regarded as certain. If decision maker acts
consistently with his or her stated preferences, the
alternatives will be chosen following the
prescriptive analysis, which are based on axioms.
Basic terminology
Objectives: An objective is an indication of
preferred direction of movement. i.e: buy a safe
house.
Attribute: It is used to measure performance in
relation to an objective, i.e: number of times this
house was stolen.
Value versus Utility: If the decision involves no
elements of risk and uncertainty for each course of
action, we will refer to this score as the value;
otherwise, we will refer to this score as a utility.
An office location problem
A small printing business must move from its existing office.
The owner is considering seven new possibilities.
However, the owner has a lot of factors to consider.
Addison Square (A)
Bilton Village (B)
Decision to move Carlisle Walk (C)
to a new office
Denver Street (D)
Elton Street (E)
Filton Village (F)
Gorton Square (G)
SMART – Simple Multi Attribute Rating Technique
The main stages are the following
1) Identify the decision maker: business owner
2) Identify the alternative course of action: 7 offices
3) Identify the attributes which are relevant to the decision
problem. Building the value tree.
4) Assign values to measure the performance of alternatives on
that attribute.
5) Determine a weight for each attribute.
6) For each alternative, take a weighted average of the values
assigned to that alternative.
7) Make a provisional decision
8) Perform sensitivity analysis to check how robust the decision is
facing the problems
3. Constructing the Value tree
Is the value tree an accurate and useful
representation of the decision maker’s concerns?
1. Completeness
2. Operationality
3. Decomposibility
4. Absence of redundancy
5. Minimum size
4. Measuring how well the option perform on
each attribute
Costs are values
Measuring the benefit attributes is a little bit
difficult.
Variables to represent those attributes, i.e.:
size.
Direct Rating
Attributes such as “comfort” or “image” will be
more difficult to find a variable which can be
quantify.
Starting with the attribute “image”: the owner is
first asked to rank the locations in terms of their
image from the most preferred to the least
preferred.
Addison Square: best location 100
Carlisle Walk: worst location 0
Direct Rating
The owner is now asked to rate the others
locations
This is a interval scale, like temperature!
Appropriate question:
Are you happy that the improvement in Image
between Carlisle Walk and Denver Street is less
preferable then that between Denver Street
and Elton Street? How much?
Bisection- Value functions
Let us now consider the attributes which can
be represented by easily quantified variables.
Bisection - Value functions
Larger office more attractiveness
OFFICE AREA (ft2) VALUE
Elton Street 1500 100
Carlisle Walk 400 0
Filton Village 400 0
The owner agrees that 700 ft2 has the midpoint, 500 ft2 is the
quarter point and 1000 ft2 represents the value 75
AREA
VALUE
700 ft2 50
500 ft2 25
1000 ft2 75
Value functions
5. Determining the weights of the attributes
If the options perform very similarly on a particular attribute, so that the
range between worst and best is small, this attribute is unlikely to be
important in the decision
5. Determining the weights of the attributes- Swing
Weights
Considering the Benefits, the owner is asked to imagine a hypothetical
office with all attributes at the least preferred level. An office that is the
greatest distance (8 miles) from the town center (not close), has the
worst position for visibility and, then, has the worst image, the smallest
size and so on.
Then, he is asked, if just one of these attributes could be moved to its
best level, which would he choose? We can improve one, taking it to the
top of the list.
The answer ranking are: (1) Closeness to customers
(2) Visibility
(3) Image
(4) Size
(5) Comfort
(6) Car-parking facilities
Determining the weights of the attributes
We can now give “closeness to customers” a weight of 100.
A swing from the worst “image” to the best “image” is
The other weight are given by comparing each attribute with
considered to be 70% as important as a swing from the worst
“closeness to customers”, giving some score (%) to the next
to the best location for “closeness to customers”
one that could be improved. Example: Visibility is 80% as
….. so “image”
important as the is assigned
swing a weight of to
in “Closeness 70.customers”.
After this, we must to normalize the values.
Calculating aggregate benefits for each
location
Addison Square
Attribute Values Weight Value
x weight
Closeness to customers100 32 3200
Visibility 60 26 1560
Image 100 23 2300
Size 75 10 750
Comfort 0 6 0
Car-parking facilities 90 3 270
8080
so aggregate benefits =
8080/100 = 80.8
6. Aggregating the benefits using additive model
7. Trading benefits against cost
Now, we will take account the costs.
The aggregate value of benefits has been plotted
against the annual costs for each of the offices.
Therefore, the only
locations which are
worth considering are A,
G and C.
7. Trading benefits against cost
The choice among the 3 offices on the
efficient frontier will depend on the relative
weight the owner gives to costs and benefits.
If he is much more concerned about benefits,
the Addison Square will be his choice.
Sensitivity analyses
It is used to examine how robust the choice of an
alternative is to changes in the figures used in
analysis.
I.e.: If turnover had a
weight of 0 this would
imply that the three
turnover attributes
would also have 0
weights, which mean
that Elton Street would
have benefits with a
value 81,7.
The axioms of the method
If the decision maker accepts these axioms, and if
he is rational then he should also accept the
preference ranking indicated by the method.
A) Decidability: We assumed that the owner was able to
decide which of 2 options he preferred.
B) Transitivity: if the owner prefers A to B and B to C, so he
prefers A to C.
C) Summation: If the owner prefers A to B and B to C, the
strength of preference of A to C must be greater than the
strength of A to B.
D) Finite upper and bounds for value: we have to assume that
the best option was not so wonderful and the worst option
was not so awful.
