Sie sind auf Seite 1von 8

24TH INTERNATIONAL CONGRESS OF THE AERONAUTICAL SCIENCES

INTEGRATION OF A SCENARIO METHOD IN AIRBUS’


TECHNOLOGY AND PROJECT EVALUATION
Matthias Meussen*, Axel Becker**
*AIRBUS, **DaimlerChrysler

Keywords: scenarios, technology, evaluation

Abstract evaluation process is described, which is


intended to improve the generation of this
For a technology evaluation that is based
information in a structured and traceable way.
on technical and economic models, quantified,
numerical (input) parameters are prerequisite.
As these parameters are projections of a (far) 2 Technology Evaluation
future they are by nature difficult to obtain and The process of technology evaluation at
subject to discussion. To deal with the wide AIRBUS is intended to put the benefit of an
variety of future developments, methods have individual technology into an overall aircraft
been used to analyze the environment and to context at a comparable basis. To cover the full
create consistent scenarios. set of effects associated with it, a cost-benefit
The paper describes the implementation of analysis for manufacturer and operator is
a modified scenario process in the technology computed. This process comprises a technical
and future projects assessment method at assessment, an economic evaluation and a risk
AIRBUS. It describes in brief the approach of analysis.
technology evaluation and risk analysis that is
based on technical and economic models. The
3 Risk Analysis
intentions of scenario use and the baseline
scenario method are outlined. The modifications Trying to obtain a prediction of the benefits
to the method are discussed and the of a technology for a far future is most obvious
implications are shown, including an example subject to uncertainty. Two major areas of risk
application. are dealt with while carrying out AIRBUS’
technology evaluation process:
• the technology inherent risk – any
1 Introduction risk that arises out of the
The development of new technologies development, industrialization and
requires long-term investments in terms of time, use of a technology in an aircraft
resources and money. To secure these • the risk of changing targets,
investments and to guide the development of requirements and environment for
technologies towards a successful the application on future aircraft
implementation into future products, AIRBUS It is the later risk that shall be tackled with
has established a process to support technology the scenario method described here.
evaluation. This process requires information on For the technology inherent risk, two
long-term business development of markets complementary approaches are established.
(social and economic environment), future First, an identification of the risks throughout
requirements and demands, and the technical the life of the technology is initiated. Even
characteristics of the technology. In this paper, a though this list of risks cannot be complete, it
scenario method as part of the technology does allow an estimation of the capability of a

1
AUTHOR 1,Fehler! Verweisquelle konnte nicht gefunden werden.

Parameter 1

Technology
Evaluation
& Viability
Parameter 2
Scenario 1
Scenario .
.
Parameter 3

Process .
.
Parameter 4

Scenario X Parameter 5

Parameter 6

Fig. 1 Scheme of Scenario Based Technology Evaluation

technology to achieve its targets. Furthermore, it qualitatively describes the interdependencies


is a starting point for an accompanying risk between the parameters. Both a-priori estimates
management. and CI-Matrix are then transferred by a
Second, the relevant parameters of a calculation method into a-posteriori
technology in the context of technical and probabilities that build the basis for multiple
economic evaluation are identified and the scenarios (method of conditional probabilities).
uncertainty of these parameters can be This scenario process, conducted with a
quantified. This is done in the form of group of experts in a workshop format, is
probability distributions derived from experts’ comprised of eight steps (Fig. 2):
discussions. Once the uncertainty has been
modeled, a Monte-Carlo-Simulation is run using Step 1: Define Focal Issue
the same models as for the deterministic In the first step, the workshop participants
technical and economic evaluation. define the focal issue to be investigated in the
The scenario process described here is used scenario process — the specific topic, its
to identify and quantify the uncertainties geographical scope, and the time horizon. This
stemming from the future environmental issue definition sets the parameters for step two,
conditions, targets and requirements for future the selection of key factors (premises and
products. From the metrics oriented scenario descriptors), as well as the nature of the scenario
process probability distributions for the scenario descriptions that are created in step five.
descriptors are derived that are used to feed the
quantitative risk analysis of the technology Step 2: Identify Premises and Descriptors
evaluation process (Fig. 1). This approach In step two, the participants identify the key
allows using the same simulation techniques for factors that will shape the focal issue. Following
technology inherent and targeting related risks. a group brainstorming session to generate an
initial list, these factors are categorized,
prioritized, and separated into premises and
4 The Conventional Scenario Process descriptors. Premises are treated as given in the
scenario process. Those factors for which
The conventional scenario approach starts
multiple outcomes are possible (e.g., energy
with a description of the relevant parameters for
prices could increase, decrease, or remain
the problem under consideration. For each of
stable), are called descriptors. They are treated
these parameters that will later characterize a
as variables in the scenario process.
scenario, a bandwidth of possible future
outcomes is chosen. The participants of a Step 3: Define Key Factors; Project
scenario process are then asked to give a-priori Descriptors
estimates of the probability of occurrence of In step three, the participants precisely
each state of parameter. In a subsequent step, a define the premises, assess their current status,
Cross-Impact-Matrix (CI-Matrix) is created that and give the reasons for their future
2
PAPER TITLE

1 2 3 4

Identify Key Define Key Factors


Define Factors Cross-Impact
Focal Question (Premises, Project Matrix
Descriptors) Descriptors

8 7 5 6

Implications Wild Card/


Recommendations, Select and Build
Sensitivity
Strategic Options Scenarios
Analysis
Early Indicators

Fig. 2 The Eight Scenario Steps

development. They also precisely define the • Scenario Selection: The goal of the
descriptors and assess their current status. After selection process is to obtain a
this, they make projections about the likelihood diverse set of two to four plausible
of each of the possible outcomes occurring and scenarios against which the
provide the reasoning behind each of these workshop participants can assess
projections. the implications and strategic
options in steps 7 and 8. While
Step 4: Perform Cross-Impact Analysis there is no fixed rule for selecting
In order to construct internally consistent the scenarios, three criteria serve as
scenarios, it is necessary to first determine if general guidelines: frequency,
and to what degree the descriptors influence one internal consistency, and variety.
another. In step four, therefore, the participants The Scenario Tool generates the
use a cross-impact matrix to examine the following data, which serve as the
interrelationship between each pair of foundation for the participants’
descriptors, quantifying the degree of influence final selection.
on a scale from strongly negative to strongly The second scenario often has a very
positive. different, though equally realistic set of
descriptor outcomes. The third and/or fourth
Step 5: Select and Build Scenarios scenarios, if chosen, generally contain elements
After analyzing the cross-impact matrix that are either surprising, unusual, or threatening
input, the STRG Scenario Tool generates data to the company. Frequently, these scenarios
that enable the workshop participants to contain a disruptive event, or wild card (see
accomplish two tasks: Step 6 below).
• Descriptor Reassessment: One goal Once the participants choose their
of the scenario process is to assess scenarios, their task is to transform them from a
how individual descriptors change list of descriptor projections into descriptions or
once their interrelationships with stories that are both realistic and accessible to
one another are taken into account
in the cross-impact matrix.
3
AUTHOR 1,Fehler! Verweisquelle konnte nicht gefunden werden.

people who did not take part in the scenario is done, may make specific recommendations
process. about future company actions. Ideally, the
participants will derive a “robust” strategy from
Step 6: Conduct Wild Card/Sensitivity the scenarios; that is, a strategy that would work
Analysis: well in any of the possible futures or at least to
Once the participants have developed their be flexible enough to adapt to any of them on
scenarios, step six is used to test how the short notice.
scenarios change when a descriptor projection
or cross-impact analysis is altered. This is 5 Modifications in the Scenario Process
known as a sensitivity analysis, for it checks the
sensitivity of the scenario to changes in its For the use of the scenarios process
internal makeup. In addition, external or described above for technology evaluation,
disruptive events can be introduced into the some of its steps have been modified, others had
scenarios (by inclusion in the cross-impact to be disconnected to either pre- or post
matrix) to see how they change the descriptor processing. Especially the last two steps (7 and
outcomes. These discontinuities, known as wild 8) are of minor importance in this context.
cards, are developments whose likelihood of To cope with the demand for agreed and
occurrence is less than 10%, but whose impact consistent parameters for technology evaluation,
on the focal issue would be very high. If the thee elements are introduced into the scenario
wild cards or new descriptor projections have a method:
surprising and/or significant impact, the • the definition of strictly metrics
participants can use them to create new oriented parameters (a quantified
scenarios. description rather than a qualitative
one),
Step 7: Implications/Early Indicators: • the split of the parameters (and the
In Step 7, the participants first construct a CI-Matrix) into global, intermediate
set of early indicators for each scenario to assist and local factors and
in monitoring which of the scenarios is actually • the transformation of the scenario
evolving in the forthcoming years. Using the parameters into probability
projection from the most important descriptors distributions for quantitative risk
in each scenario, the participants identify one or analysis.
more indicators of this projection and establish a
threshold measure, which, if crossed, would 5.1 Quantifiable Descriptors
indicate that the projection is becoming a The technology evaluation at AIRBUS is
reality. From these early indicators, it is possible based on technical and economic simulation of
to construct an “early warning system.” In the the effects of the technology at aircraft level.
other component of Step 7, the participants The underlying models require of course the
assess the implications of each scenario for their input of the parameters describing a scenario in
industry, their company strategy, or a specific scalar variables or in the case of quantitative
product. In addition to the scenario-specific risk analysis in the form of probability
implications, implications that apply across distributions.
several or all scenarios are also of significance. As traditional scenario processes do not
necessarily need this precision in the definition
Step 8: Assess Strategic Options; Make of their descriptors, a special focus has to be set
Recommendations on the generation of quantified descriptors in
In the final step of the scenario process, this approach in step 3 of the above process.
participants assess the strategic options Once having identified the relevant
available to the company and, depending upon premises and descriptors in step 2, it is
the larger context in which the scenario process necessary to translate these into representative
4
PAPER TITLE

metrics. These metrics have to fulfill multiple for technology evaluation, the quantification
criteria. They should be continuous, explicit and allows to track the evolution of the perception
well defined. over time and a constant update with newly
To illustrate the generation of quantifiable available statistical data.
descriptors an example shall be given. In a For the task of transformation of
scenario process, the participants identified the qualitative to quantitative descriptors, three
descriptor “International Political Stability” to different techniques have been used and found
take into account the state of balance of political useful:
power. It embraces the political situation that • deriving the descriptor directly
changes in the frame of economic and politic from the calling models or
transformation processes between different programs,
groups of interest. • asking experts in the relevant field
The metrics identified for this descriptor for their input,
• searching the Internet for the
180

Political Conflicts
keywords and analyzing the
160
associated metrics.
140
As this procedure requires considerable
120
time, most of the parameters have to be
100 prepared prior to the workshops and are offered
80 to the participants to choose from. This way of
60 working puts additional load on the preparation
40 team and has to be performed quite carefully in
20
order not to bias the results. Another solution
0
would be to break after step 2 and to perform
1995 1996 1997

violent conflicts
1998 1999 2000 2001

nonviolent conflicts
2002
the necessary analysis offline.

Fig.3 Average No. of Conflicts Worldwide p.a. [1] 5.2 Cascade approach
For the method described here, the
describing parameters are grouped in three
was the number of political conflicts per year. areas: the macro factors are related to the global
Statistical data was available for the last decade socio economic environment, the meso factors
(Fig. 3, [1]), so a baseline definition could be cover air transport in general and the micro
derived including a “state-of-the-art” statement. factors are specific to aircraft requirements and
In the course of the process, qualitative technology impacts (Fig. 4).
projections (“more stability”, “as today”, “less It is assumed that only the “upper level”
stability”) could be transformed into quantified environment factors will influence the “lower
metrics. level” factors. This goes along with a
From this statement of the current requirement driven perception of the air
situation, the definition and three projections transport industry and the aircraft manufacture.
were derived: It is recognized that different concepts of the
• decreasing no. of conflicts (100 – influence of air transport on the economy exist
140 conflicts per year) (e.g. new routes generating additional economic
growth) but these are considered to be of minor
• stagnating no. of conflicts (140 –
interest in more global scenarios. Nevertheless,
180 conflicts per year)
they might be of importance if a more local
• increasing no. of conflicts (180 –
focus for the problem definition is chosen.
220 conflicts per year)
As the CI analysis becomes very time
Even though the bandwidth for the
consuming with the number of parameters
projections does not directly go into the models
neglecting one third of all possible
5
AUTHOR 1,Fehler! Verweisquelle konnte nicht gefunden werden.

Fig. 4 Cascade of Scenario Factors

interdependencies results in a considerable the micro environment it is obvious that


reduction of effort. different scenario processes have to be
The concept of splitting or cascading the conducted. The cascades approach allows to
scenario descriptors into global, intermediate integrate the more detailed “lower level”
and local factors (or macro, meso and micro approaches into a single, common overall
factors) takes into account the fact that in scenario picture.
general, it is difficult to unite experts for all
aspects of the complex scenarios at one time 5.3 Evaluation of Cross-Impact Matrix
and at one place for discussion. It was obvious A generation of a CI-Matrix that can be
at several processes that political economists are altered and updated throughout the process of
not too comfortable in the generation of scenario generation can be looked upon as a
projections on aircraft approach speed as basic model of interdependencies between the
aeronautical engineers are not with GDP factors that describe a scenario.
prediction. In order to construct internally consistent
The building-block structure also allows scenarios, its first necessary to determine if and
integrating already existing scenarios (e.g. to what degree the variables influence one
Europe’s Vision 2020 on air transport) into a another. The experts use a Cross-Impact-Matrix
more detailed and still consistent view on to examine the interrelationships between each
specific aspects of technology and project pair of variables, quantifying the degree of
evaluation. It can be assumed that the global influence on a scale from strongly negative to
view from the macro environment may remain strongly positive.
constant while meso and specifically micro Even though it is based on more qualitative
environment factors are subject to adaptation to relations, it is shown that it gives a good first
the technologies under consideration. Within a rationale for the selection of relevant drivers of
single scenario process the number of factors is future developments and the stability or
limited to a maximum around 30 for practical robustness of the scenarios derived from. Key
reasons. Considering the wide set of potential question for cross-impact analysis:
requirements for a technology evaluation out of • How will the probability of
occurrence of the respective state of
6
PAPER TITLE

(row-) variable Y be changed if the Example: Variable “Payload“ with the indicator „ No. of passengers“

respective state of (column-) Alternative


Projections
a-priori-probability
(scenario tool input)
a-posteriori- probability
(scenario tool output)
variable X occurs? A: 450 - 550 Pax 40 % 32 %

A scenario is always calculated under the B: 350 - 450 Pax 40 % 55 %


C: 250 - 350 Pax 20 % 13 %
assumption that one certain future projection of
a variable occurs, i.e. that the probability of this
40 % 40 % 40 %
variable is equal to 100 percent. From this 55 % 55 %
starting point the probabilities of all other 20 % 40 % 40 %
20 %
20 %
32 %
20 %
32 %

variables will be recalculated according to the 13 % 13 %

250 350 450 550 250 350 450 550 250 350 450 550
matrix values and the scenario tool generates the
entire range of possible future scenarios. Each Fig. 6 Probability Distributions for Descriptors
of these skeletal scenarios is comprised of
different combinations of descriptor outcomes.
In conventional scenario projects (Fig. 5a)
single scenarios or scenario cluster were robustness can be made either way and may
generated to have a sufficient spectrum of complement each other.
possible future outcomes. Technology
evaluation was only conducted on single 5.4 Descriptor Bandwidths for Risk Analysis
incidents and situations. Once having spectrums for the projections
In this continuative approach (Fig. 5b) at hand, the probability distributions can be used
additional statements to the entire scenario to initiate a numerical risk analysis. Describing
spectrum and certain factors are computed (in the uncertainty of some boundary conditions in
the form of probability distribution). the economic evaluation and assessing their
effects with Monte-Carlo Analysis then gives an
indication of the robustness of a technology
under uncertain future targets.

6 Example of Application
To cope with the challenges of future air
transport and to identify promising technologies
and concepts some “out-of-the-box” thinking
a b becomes necessary. For this exercise, AIRBUS
has created a set of unconventional
Fig. 5 Scenario Clusters (a) and Scenario Spectrum (b) configurations to explore capabilities and to
meet more demanding targets [2].
Different scenarios were created that were
From the results of the scenario tool a set up under different paradigms to drive certain
single probability distribution can be derived for requirements and to open the associated design
every variable to isolate the start input for the spaces. To assess each configuration in the
risk assessment of product and technology relevant context, not only the primary drivers
evaluation (Fig. 6). had to be determined but a consistent
Using the same process for either environment (comparable but distinguished
generation of single scenarios and spectrums of from the current evaluation models) needed to
projections allows to analyze in the more be created. These scenarios were then used to
abstract way of quantitative risk analysis and to compare different configurations (Fig. 7).
compare that analysis with the well defined Without going into details of the evaluation as
consistent scenarios. An assessment of such, it can be seen that the effects of different

7
AUTHOR 1,Fehler! Verweisquelle konnte nicht gefunden werden.

The Classic Performer The Money Buster The Proactive Green

B Classic Scenario B Economic Scenario B


Green Scenario
e e e
n n n
e e e
f f f
i i I
t t t

Fig. 7 Comparison of Aircraft Concepts in Different Scenarios

scenarios had a significant impact on the benefit is available. The communication of the rationale
assessment of the technologies. and trade factors used for optimization in
different scenarios is improved. Still, some
6 Conclusion aspects of the interpretation of mathematical
formulation and interpretation of probable
Due to the long term of projections results are subject to further investigation.
necessary and the high degree of uncertainty
associated, the input parameters required to
perform the technology evaluation are difficult 7 References
to obtain. Even when it is not possible to [1] Heidelberg Institute of International Conflict
improve the accuracy of the predictions, a Research HIIK, 2003
(http://www.hiik.de/de/index_d.htm)
structured and repeatable, cascaded approach
[2] Y. Vigneron, “Commercial Aircraft for the 21st
using a metrics oriented scenario process is a Century: A380 and Beyond”, AIAA/ICAS Air &
way to get a higher quality of analysis as well as Space Symposium and Exposition, Dayton 2003
an enhanced common understanding.
The use of scenario processes in
technology and project evaluation has to go
along with methodologies in decision and policy
making (e.g. Vision 2020, ASTERA). Breaking
up the process in several cascaded environments
and generating an expandable CI Matrix
supports this approach.
A further issue is the necessity of an
improved understanding of the effects of the
scenario uncertainty on the deterministic
technology evaluation. Here is the challenge on
the interpretation of the results.
With the here described method of the
integration of a scenario process into technology
evaluation a procedure to better understand and
more efficiently guide technology development
8