FINAL REPORT

COMPLEX DECISION OF BUYING OR NOT BUYING NEW

AIRCRAFT IN AIRLINE COMPANY

IE 5203 DECISION ANALYSIS

ADITYA NUGROHO
HT083276E

DEPARTMENT OF INDUSTRIAL ENGINEERING

NATIONAL UNIVERSITY OF SINGAPORE

2010
 Department of Industrial Engineering
IE 5203 Decision Analysis

EXECUTIVE SUMMARY

This term paper tries to present a real world problem in Decision Analysis. An airline
company ABC in Indonesia is plan to develop Jakarta-Singapore route network on short haul
flight within ASEAN countries. The present operation makes use of Boeing 737-300. The
company must decide whether to maintain present aircraft or buy a new aircraft that will
increase its capacity.

Value-focused thinking was done to identify values that matter to the decision maker. Thus,
alternatives were generated. To consider the monetary gain of each alternative in term of
NPV, one pass through DA cycle using DPL programme was done. Uncertainties of future
income, maintenance cost, and the salvage value of the aircraft was accounted for.

To consider other intangible factors, Analytical Hierarchy Process (AHP) was done using
Expert Choice software. Criteria and subcriteria to determine the best alternative were laid
down in Tree Hierarchy. The weightage of each criteria and subcriteria were calculated using
pairwise comparison technique. All alternatives were then compared to each other. Best
alternative was obtained.

2
 Department of Industrial Engineering
IE 5203 Decision Analysis

Table of Contents

1.0 INTRODUCTION AND PROBLEM DESCRIPTION .................................................. 4

2.0 FRAMING AND FORMULATION............................................................................... 4
2.1 Framing........................................................................................................................ 4
2.2 Formulation ................................................................................................................. 6
2.3 Influence diagram and generic decision tree ............................................................... 7
3.0 DETERMINISTIC ANALYSIS ..................................................................................... 8
4.0 PROBABILISTIC ANALYSIS ...................................................................................... 9
5.0 MODEL APPRAISAL.................................................................................................. 12
6.0 ANALYTIC HIERARCHY PROCESS MODEL......................................................... 13
7.0 RECOMMENDATIONS AND CONCLUSIONS........................................................ 16

Figures and Tables

Table 1 Alternatives comparison................................................................................................ 5

Table 2 Future uncertainties....................................................................................................... 6
Table 3 Value model of investment decision ............................................................................. 7

Figure 1 Influence diagram and generic decision tree of the investment model........................ 7
Figure 2 Tornado diagram.......................................................................................................... 9
Figure 3 PMF for the aleatory variables..................................................................................... 9
Figure 4 Influence diagram of probabilistic analysis ............................................................... 10
Figure 5 Value node definition................................................................................................. 11
Figure 6 Risk profile ................................................................................................................ 11
Figure 7 DPL Optimal decision policy .................................................................................... 12
Figure 8 Value of information and control............................................................................... 12
Figure 9 Sensitivity analysis of pax demand node (rainbow diagram) .................................... 13
Figure 10 AHP Hierarchy ........................................................................................................ 14
Figure 11 Hierarchy and synthesis with respect to Goal.......................................................... 14
Figure 12 Sensitivity analysis with respect to main criterion .................................................. 15
Figure 13 Sensitivity analysis with respect to intangible criterion .......................................... 16

3
 Department of Industrial Engineering
IE 5203 Decision Analysis

1.0 INTRODUCTION AND PROBLEM DESCRIPTION

The airline company ABC in Indonesia is plan to develop Jakarta-Singapore route network on
short haul flight within ASEAN countries. The present operation makes use of Boeing 737-
300. According to air transportation bureau, it is expected that the demand for air passenger
would be 65,000 pax per annum. The company is now considering to invest in a new aircraft
with huge capacity but suitable with annual passenger demand. By this new aircraft the airline
could carrying more passenger for international route.

Making this decision is hard because of the complexity by the many choices of aircraft
technologies: the capacity, the reliability of the technology, the market value, after sale
service, and so on. However, only some alternatives will be considered in this paper, because
of the limited capability of the trial version software used.

In addition there are also many objectives to be fulfilled, such tangible and intangible factors.
Considering investment in new aircraft not only the expected payback period and future
income but also the ease and safety of using the technology, maintenance, and the possibility
of business expansion.Thus, there’s another reason why making this decision is hard, that is
because the future is full of uncertainties. This could come from the economic situation in
ASEAN countries that will affect the the demand annual passenger.

2.0 FRAMING AND FORMULATION

Value-focused thinking is used to formulate the problem. Values & objectives will be
identified and structured. Influence diagram and decision tree will be used to structure the
elements of the decision situation.

2.1 Framing

a. Target
 The company needs to earn money
 The company wants to become market leader
 The company wants to expand the business

b. Alternatives
 Maintain old aircraft: maintain the current profit, limited by the aircraft capacity
 Buy new aircraft: invest a sum of money, monetary gain in the future, and increase
capacity.

4
 Department of Industrial Engineering
IE 5203 Decision Analysis

 Alternatives of the new aircraft (2 alternatives have been chosen):

- Boeing 737-900ER
- Airbus 330-200 series
Comparison of the two alternatives and estimated value of the investment from the
aircraft manufacturer is shown in Table 1.

c. Problems
 The old aircraft is giving problem of limited capacity to carrying more passenger

Table 1 Alternatives comparison

Boeing 737-900ER Airbus 330-200 series

Manufacturer Boeing The European Aeronautic
Defence and Space Company
Max seat capacity 215 passengers in a single- 293 passengers in a two class
class
Max take off weight Weights 187,700 lb (85,130 Weights 233,000kg
and max range flies kg), Flies up to 3,265 nautical (513,670lb), Flies up to
miles (6,045 km) 11,850km (6400nm)
Maximum fuel capacity 7,837 U.S. gal (29,660 L) 36,750 US gal. (139,100
Litres)
Reputation of Established, pioneer ing Established consortium with
manufacturer medium aircraft size European company, very
active in research
Service Worldwide including Worldwide including
Singapore Singapore
Aircraft safety precision The -900ER is a performance EASA determined that the
category D (The -800 is A330-200F does not present
performance category C) unique characteristics that
require flight evaluation.
Estimated price US$76.0 - 87.0 (millions) US$170.9 - $200.8 (millions)

d. Uncertainties
It is predicted that the future condition is full of uncertainties.
 The monetary gain by investing in new aircraft will be greatly affected by future
condition which is determined by the economical situation in ASEAN countries.
 The consequences of buy new or maintain current aircraft would is shown in Table 2

5
 Department of Industrial Engineering
IE 5203 Decision Analysis

Table 2 Future uncertainties

Economic condition Maintain New investment

Better Limited by capacity Carrying more passenger,
more profitable
Stable Stable Possibility to expand the
international routes
Worse Stay on Lose money

e. Constraints
 Investment in new aircraft required huge initial cost and need credit service from the
Bank.

f. Strategic objective
 Value that is absolutely fundamental : more profit

2.2 Formulation

a. Decision variables
 Buy Boeing 737-900ER
 Buy Airbus 330-200 series
 Maintain old aircraft Boeing 737-300

b. State or system variables

 Demand (passenger/annual): low, base, high
 Annual maintenance cost (million /year): low, base, high
 Salvage value (million): low, base, high

c. Values and preferences

 NPV (Net Present Value) of net income at MARR = 10%
 Study period at 15 years
 It is consder that the company will be take the Risk Neutral

6
 Department of Industrial Engineering
IE 5203 Decision Analysis

2.3 Influence diagram and generic decision tree

O&M Salvage
Demand Cost Value

Investment
NPV
Choice

N
MARR Gross
profit

Investment O&M Salvage

Demand
Choice Cost Value

Buy Boeing 737-900ER Low Low Low

Buy Airbus 330-200 series Nominal Nominal Nominal

Maintain old aircraft Boeing 737-300 High High High

Figure 1 Influence diagram and generic decision tree of the investment model

Table 3 Value model of investment decision

Investment Decision Model Short Haul Flight

Decisions
Buy B737 900ER
Buy A330-200
Maintain B737-300

Inputs Sensitivity range

Low Base High
Annual passenger demand 65000 35000 65000 95000
O&M cost B737 900ER $19,392,400 $9,696,200 $19,392,400 $29,088,600
O&M cost A330-200 $26,853,353 $13,426,677 $26,853,353 $40,280,030
O&M cost B737-300 $17,636,000 $8,818,000 $17,636,000 $26,454,000
Salvage value B737 900ER $35,000,000 $17,500,000 $35,000,000 $52,500,000
Salvage value A330-200 $80,000,000 $40,000,000 $80,000,000 $120,000,000
Salvage value B737-300 $15,000,000 $7,500,000 $15,000,000 $22,500,000

Fixed data

7
 Department of Industrial Engineering
IE 5203 Decision Analysis

Study period 15
MARR 10%
Initial cost of B737 900ER $90,000,000
Initial cost A330-200 $160,000,000
Gross operating cargo profit/pax
B737 900ER $300
A330-200 $500
B737-300 $100
Air fare; per passenger $250

Intermediate values
Uniform series PWF (P/A,10%,20) 7.6061
Single payment PWF (P/F,10%,20) 0.2394

Value model
Strategy
1. NPV Buy B 737-900ER $42,795,928
2. NPV Buy A 330-200 $25,698,999
3. NPV Maintain B 737-300 $42,488,371

3.0 DETERMINISTIC ANALYSIS

From the above value invesment decision model thus we will consider deterministic analysis.
The purpose of deterministic analysis is to choose sensitive variables, which influence the
result greatly with just a little change of value. These variables are called aleatory variables
and will be considered further in the probabilistic analysis, while other insensitive variables
will be set to the base value.

By using the Sensit software, the state system variables are calculated and deterministic
analysis is done and the result is shown as the Tornado Diagram (Figure 2). From the tornado
diagram, we can analyzed as follows:

 As the business operating of airline is depend on passenger demand. The model shows
that the annual passenger demand influence the total present worth of NPV income
greatly.
 Secondly the range of operation and maintenance cost of aircraft also influence the
value of NPV income although not as great as the annual passenger demand.
 The fluctuation of salvage value of aircraft costs don’t change the value of present
worth significantly.

As a conclusion of this deterministic analysis, it is decided that the variable of annual

passenger demand and O&M cost of aircraft will be treated as aleatory variables, while other
variables will be set at the base values for subsequent analysis.

8
 Department of Industrial Engineering
IE 5203 Decision Analysis

Annual passenger demand 35000 95000

O&M cost A330-200 $40,280,030 $13,426,677

O&M cost B737 900ER $29,088,600 $9,696,200

O&M cost B737-300 $26,454,000 $8,818,000

Salvage value A330-200 $40,000,000 $120,000,000

Salvage value B737 900ER $17,500,000 $52,500,000

Salvage value B737-300 $7,500,000 $22,500,000

-$400.000.000 -$200.000.000 $0 $200.000.000 $400.000.000 $600.000.000

Total NPV

Figure 2 Tornado diagram

4.0 PROBABILISTIC ANALYSIS

In this section the pdf (probability distribution function) of each aleatory variables will be
assessed. Using the Pearson-Tukey three-point quick approximation method, the CDF
(cumulative distribution frequencies) derived is discretized, resulting in the PMF (Probability
Mass Function) for each aleatory variables. Pearson and Tukey suggested using the 5, 50, and
95 percentiles and in this case the branch probabilities of the approximate pmf are [0.185,
0.630, 0.185]. Following figure represent the pmf of each aleatory variables.

Figure 3 PMF for the aleatory variables

9
 Department of Industrial Engineering
IE 5203 Decision Analysis

The probability analysis is done using the DPL program by considering aleatory variables in
influence diagram (see Figure 4). Some snapshots and results from the DPL analysis are
shown in Figure 5 and 6. The optimal decision is shown in Figure 7.

B739
B733 OM Pax
B739 A332
OM demand
SV OM

A332
SV
B733
SV Investment
NPV choice
MARR

B7379 A3302 NPV

I I B739
A3302
P
NPV
B7379 A332
P Airfare

NPV
B7373 B733
P

Figure 4 Influence diagram of probabilistic analysis

10
 Department of Industrial Engineering
IE 5203 Decision Analysis

Figure 5 Value node definition

Figure 6 Risk profile

As Figure 6 shows, the alternative exhibits non stochastic dominance over the other
alternatives. Note, however, in spite of greater variability in parameters associated with the
Buy Airbus option, the corresponding risk profile is relatively tight.

Therefore, by considering the monetary gain and future income therefore the optimal decision
policy of airline company is to Buy Boeing 737-900ER for operating short haul route for
Jakarta-Singapore (see Figure 7).

11
 Department of Industrial Engineering
IE 5203 Decision Analysis

Figure 7 DPL Optimal decision policy

5.0 MODEL APPRAISAL

In this section I would like to discuss analysis of value information and control. Value of
information is the amount a decision maker would be willing to pay for information prior to
making a decision. The value of control is a quantitiative measure of the value of controlling
the outcome of an uncertainty variable.

Figure 8 Value of information and control

As shown in Figure 8 the manager of airline company can run a value of control or value of
information diagram to see which nodes most directly affect the outcomes. As we have four
uncertainty nodes in our model, the graphs show that the passenger demand is important
nodes. In addition, the value of control shows the amount of risk that could be reduced given
perfect control over each probabilistic node, and that it is clear that passenger demand would
be the most important variable for risk managers to control. Admittedly, this is a basic

12
 Department of Industrial Engineering
IE 5203 Decision Analysis

example, but with a more complex model, analysts could determine which nodes are
positively or negatively affect the outcomes and which uncertainties are most important.

Figure 9 Sensitivity analysis of pax demand node (rainbow diagram)

By using DPL software is allow to easily perform sensitivity analysis on key model
assumptions. From the value of information and control above, the Expected Value of
passenger demand was highly. We can generate sensitivity analysis such as rainbow
diagrams. The rainbow diagram (Figure 9) shows the decision changes as our assumption
about the nominal value of passenger demand increases. The different shaded regions
represent different decisions. In conclusion the company should control the pax demand
variable which would be affect the great outcomes of the model.

6.0 ANALYTIC HIERARCHY PROCESS MODEL

In this section considering decision to buy aircraft is not only based on policy tree of NPV,
however there is exist tangible and intangible factors should take into account. Therefore in
the AHP analysis these factors will further discuss. In AHP model NPV value for each
alternatives will be use to calculate preference. In example the preference for Buying Boeing
737-900ER over the Boeing 737-300 will be 52,067,548/42,488,595=1.225

As can bee seen in Figure 9 by using the AHP hierarchy, the set of main criteria (tangible and
intangible) has been considered in second level hierarchy. Each main criteria is decomposed

13
 Department of Industrial Engineering
IE 5203 Decision Analysis

into several subcriteria. Each criteria’s and subcriteria’s weightage is obtained by pairwise
comparison technique. There are 3 alternatives to be considered. AHP analysis is done using
Expert Choice 11.5. The results of expert choice is shown in following figure.

Maximize
benefits

Tangible Intangible

After sale
Safety Manufacture
NPV Tech nology service
precision reputation
assurance

Maintain
Buy Boeing Buy Airbus
Boeing 737-
737-900 ER 330-200
300

Figure 10 AHP Hierarchy

Figure 11 Hierarchy and synthesis with respect to Goal

14
 Department of Industrial Engineering
IE 5203 Decision Analysis

Figure 12 Sensitivity analysis with respect to main criterion

As shown in Figure 11 by selecting pairwise graphical comparisons in Choice Expert,

therefore the set alternatives with respect to goal can be obtained. Results of Choice Expert
shows that the best alternative is still to buy Boeing 737-900 ER instead of maintain current
aircraft. The overall inconsistency is good, because the management is not willing to change
their judgement in the pairwise comparison. But since the chosen alternative’s priority is not
far exceeding the other two alternatives, the decision is more sensitive.

After considering synthesis of alternatives, sensitivity analysis is done to see how the changes
of main criteria’s and subcriteria’s weightage affect the ranking of the alternatives. Expert
Choice is able to generate 4 kinds of graph to do the analysis (see Figure 12). As can be seen
in the dynamic sensitivity result, the chosen alternative (Buy Boeing 737-900ER) is prefered
against the other two alternatives in tangible criterion. Under intangible criteria, alternative
Buying Boeing 737-900ER is still prefered.This means rank reversal will not occur if the
weightage of the main criteria.

15
 Department of Industrial Engineering
IE 5203 Decision Analysis

Figure 13 Sensitivity analysis with respect to intangible criterion

As can be seen from the above figure 13, under intangible criteria, preference is sensitive.
Preference will change if we change the subcriteria’s weightage. As example, Airbus 330-200
with respect to safety precision subcriteria dominates other 2 alternatives. However,in overall
alternative Buy SLA Boeing 737-900ER still dominates over the other 2 alternatives.

7.0 RECOMMENDATIONS AND CONCLUSIONS

Based on DPL and Expert choice results, some recommendations for investment choice as
follows:
 Buy Boeing 737-900ER in considering monetary gain
 However, in the case weightage for intangible criteria is changed to more than 50%, the
preference will be to buy Airbus 330-200

In regards with Decision Analysis cycle following recommendations are needed for 2nd pass
DA cycle :
 Risk neutrality was assumed and the dollar value is used to obtain the expected value
of optimal decision in the 1st pass. So for the 2nd pass the risk attitude can be
considered and the utility function of the decision maker can be derived.

16
 Department of Industrial Engineering
IE 5203 Decision Analysis

 Pearson-Tukey three-point short cut approximation method was used in the 1st pass,
so in the 2nd pass a more accurate method can be used to discretize aleatory variables
such as using equal area method or fitting a certain distribution to data.
 MARR, study period was set at 10% and 15 years respectively. In the 2nd pass
analysis can be done whether the changes of these parameters are significant enough
to treat them as aleatory variables.
 The use of AHP implies that an Additive Value Function is used, hence the Additive
Independence holds. In reality it might not be the case, so investigation into this must
be done, and Multi Attribute Utility Function could be used instead.
 In calculating NPV, the means of financing this investment has not been considered. If
financing comes from outside source and there is cash outflow to pay the interest, then
it’s more appropriate to calculate the present worth of cash flow after taxes.

In conclusions, this project presented a real world problem in Decision Analysis. It manages
to combine tangible and intangible objectives to obtain the best alternative It manages to take
into account the risk of future value of demand and the uncertainty of maintenance cost and
salvage value of the aircraft.

REFERENCES

R.T. Clemen and T. Reilly, Making hard decisions with DecisionTools. Duxbury Thomson
Learning, 2001.
T.L. Saaty, The Analytic Hierarchy Process, McGraw Hill, New York, 1980
K.L. Poh, IE5203 Lecture Notes, 2010 Edition.
Applied Decision Analyis LLC, DPL 4.0: Professional Decision Analysis Software -
Academic Edition, Duxbury, 1998.

Source of data information

 http://www.airbus.com/store/mm_repository/pdf/att00011726/media_object_file_ListPric
es2008.pdf
 http://www.boeing.com/commercial/prices/
 http://www.boeing.com/commercial/737family/pf/pf_900ER_fact.html
 http://www.airbus.com/en/aircraftfamilies/a330a340/a330-200
 http://www.icao.int/icao/en/ro/allpirg/allpirg4/wp28app.pdf
 Using an operating cost model to analyse the selection of aircraft type on short-haul
routes http://www.saice.org.za/Portals/0/pdf/journal/vol48-2-2006/vol48_n2_a.pdf

17