You are on page 1of 48

Decision Making

Evaluation of Projects

• Rarely would one project emerge as the best on all the chosen criteria.

• If this happens it is the dominant project and should clearly be chosen.

• In general there would be a set of non-dominated projects the choice out of which is not
normally easy.

• Management priorities to various criteria could help decision making.


CRITERIA

TANGIBLE INTANGIBLE
(measurable in Rs, percentage, years..) ( not measurable on a well defined scale)

Not measurable

(score on a subjective scale 0-9, …)


Decision Matrix

Criteria / C1 C2 C3 … Cn
Projects …
P1 X11 X12 X13 X1n S1

P2 X21 X22 X23 X2n S2

Pm Xm1 Xm2 Xm3 Xmn Sm


Non dominated solutions

Project P1 is clearly better than p2 ,


P6 is better than P2, P3, P4 and P5
P8 better than P3, P4, P5 and P7
But P8 does not dominate P1, P2 and P6
P1, P6 and P8 are not dominated by another points.( All other projects are dominated by at least one of them)
Pareto Optimal (Non dominated) solutions

But difficult to make choice between P1 and P6


The Preferred Solution

• The dominant solution, if one exists.

• The Pareto optimal or the non dominated solution set .

• The selection from out of the non dominated solutions will involve tradeoffs and
will be governed by priorities or weightages to different criteria.
Weightages to criteria

The priorities or weights to the different criteria may be obtained by:

• Mutual consultations or opinion polls.

• Pair wise comparison between criteria.

• Establishing a hierarchy of priorities and using AHP (developed by Saaty)


Criteria for Project Selection (Worst - Best)
Performance of competing projects
Criteria
C1 C2 C3 C4 C5 C6
P1 10 40 8 High Good V.Good
P2 6 25 4 V.High Medium Good
P3 8 30 10 Low V.Good Medium
P4 3 10 2 Medium Poor V.Poor
P5 2 20 2 V.Low V.Poor Poor
MCDM

• TOPSIS (Technique for Order Preference by Similarity to the Ideal Solution)

• SAW (Simple Additive Weighting)

• AHP (The Analytical Hierarchy Process)

• SMART (The Simple Multi Attribute Rating Technique )

• ANP (Analytic network process)


Important terms…
• Alternatives – These are the options which are to be evaluated for
selection of the best.

 Criteria or Attributes – These will impact the selection of


alternatives.
 Completeness: It is important to ensure that all of the
important criteria are included.
 Redundancy: In principle, criteria that have been judged
relatively unimportant or to be duplicates should be
removed at a very early stage.

• Weights – These estimates relative importance of criteria.


• Each attribute is given certain points on 0-10 or 0-100 rating
scale by a team of experts or decision makers.
TOPSIS

• The Technique for Order of Preference by Similarity to Ideal Solution (TOPSIS)

• It is a multi-criteria decision analysis method, which was originally developed by Hwang


and Yoon in 1981 with further developments by Yoon in 1987, and Hwang, Lai and Liu in
1993.

In this method two alternatives are hypothesized:

• Ideal alternative: One which has the best attributes values (i.e. max. benefit attributes and
min. cost attributes)
• Negative ideal alternative: One which has the worst attribute values. (i.e. min. benefit
attributes and max. cost attributes)

TOPSIS selects the alternative that is the closest to the ideal solution and farthest from
negative ideal solution.
Steps involved in TOPSIS …
• Step 1 – Standardize the decision matrix
• This step transforms various attribute dimensions into non-dimensional
attributes, which allows comparisons across criteria
• Step 2 - Construct weighted standardized decision matrix by multiplying attributes
weight to each rating.
• Step 3 – Determine ideal solution and negative ideal solution
• Step 4 – Determine separation from ideal solution. Si*
• Step 5 – Determine separation from negative ideal solution. Si-
• Step 6 – Determine relative closeness to ideal solution.
Fighter Aircraft Selection
(An Example)
Fighter Aircraft Example (Decision Matrix)

X4 is the cost criteria, others are benefit criteria


Scale for Intangibles
Technique for Order Preference using Similarity to Ideal Solution (TOPSIS)
Normalized Decision Matrix
Step 2: Obtain the normalized decision matrix, R, using the relationship
Weighted Decision Matrix

Step 3: Obtain the weighted decision matrix V by multiplying each column of R by the
corresponding weight.
Ideal and Negative Ideal solution

• Step 4: Obtain the ideal (A*) and the negative ideal (A-) solutions from the
weighted decision matrix V.

A* = (0.1168, 0.0659, 0.0531, 0.0414, 0.1347, 0.2012)

A- = (0.0841, 0.0366, 0.0455, 0.0598, 0.0577, 0.1118)


Seperation measures

• Step 5: Compute the separation measures from the ideal (Si*) and the negative
ideal (Si-) solutions for all alternatives, i=1,...,m.

Si* = Sq root ( sum of squares for j=1,...,n of (vij-vj*))

Si- = Sq root (sum of squares for j=1,...,n of (vij-vj-))


Values of Seperation measures

Separation measure from


Ideal Solution Negative Ideal Solution

S1*=0.0545 S1- = 0.0983


S2*=0.1197 S2- = 0.0983

S3*=0.0580 S3-= 0.0920


S4*=0.1009 S4-=0.0458
Relative closeness to Ideal solution

• Step 6: For each alternative determine the relative closeness to the ideal solution ( Ci*, i=1,...,m) as
Ci*= Si- / (Si*+Si-)
Relative closeness values

• C1* = 0.643

• C2* = 0.268

• C3* = 0.613

• C4* = 0.312
(Notice that the closeness rating is a number between 0 and 1, with 0 being the worst possible
and 1 the best possible solution)
Rank the preference order

• Step 7: Determine the preference order by arranging the alternatives in the descending
order of Ci*, i=1,...,m.

Thus the ranks for the alternatives in the fighter aircraft selection problem using TOPSIS
emerge as

• A1, A3, A4, A2


Example - TOPSIS Selection of Best Mobile phone
Using 5 point scale for looks
Fighter Aircraft Selection
(An Example)--- for SAW
Fighter Aircraft Example (Decision Matrix)

X4 is the cost criteria, others are benefit criteria


Scale for Intangibles
Simple Additive Weighting (SAW)

• Step 1: Obtain the decision matrix after converting intangibles to


numbers
Normalization

• Step 2: Obtain the normalized decision matrix R (rij, i= 1,..., m: j=1,...,n)


using

rij = xij/xj*, if the jth criterion is a benefit criterion, &

rij = xj*/xij, if the jth criterion is a cost criterion


Final Scores

• Step 3: Using the weights for the different criteria obtain the weighted score for
each alternative using the normalized decision matrix

• Step 4: Based on the final scores, rank the alternatives for a decision by the
decision maker
Normalized Decision Matrix

W (0.2, 0.1, 0.1, 0.1, 0.2, 0.3)

Ranking of alternatives with SAW : A3, A1, A4, A2


SAW and TOPSIS ( A Comparison)

• Ranking obtained using two Multi attribute decision making techniques need not
to be identical.

• For Fighter Aircraft selection Project

• SAW gave a ranking A3, A1, A4, A2

• TOPSIS gave a ranking A1, A3, A4, A2

You might also like