Professional Documents
Culture Documents
Series Editor
Rajkumar Roy
Steen Leleur
123
Steen Leleur
Department of Transport
Technical University of Denmark
Building 115, Bygningstorvet
Kgs. Lyngby 2800
Denmark
ISSN 1619-5736
ISBN 978-1-4471-2490-0
DOI 10.1007/978-1-4471-2491-7
e-ISBN 978-1-4471-2491-7
Preface
Complex Strategic Choices sets focus upon providing deliberate and methodical
support for decision-makers engaged in strategic decision making. The book aims
especially at shedding light on decision support under conditions that are uncertain
and complex. In this respect systemic planning (SP) will be introduced as a new
approachthe SP approachto deal with future-oriented, often long-term, actions
which can improve our capability of meeting present-day strategic challenges.
The book is aimed at people with an interest in new possibilities for supporting
decision making. This group of people includes managers and corporate employees
and also students and researchers of various kinds and for that matter all people
seeking to come to grips with the challenge of what we may term deliberate, informed
change. Known in some contexts as decision engineering (DE), the book will show
that DE is embedded into and dependent on a plethora of knowledge types which
range from applied mathematics known as operations research via economics and
organisation theory to social science issues. However, throughout the book focus will
be maintained on formulating a framework of practical advice. Stated quite simply,
this practice framework will aim at supporting analysts and decision-makers in
achieving better complex strategic decisions.
Normally a book is read in a linear way, which in this case means from Chap. 1
to Chap. 8. Such a reading of the book is based on the perception that it can be an
advantage that the theory is in place when the new methodology is presented as a
coherent process-and-methodology framework in Chaps. 7 and 8. However, it is
my experience from teaching planning and appraisal courses at the Technical
University of Denmark that theoretical issues appear to be more relevant and
interesting if the students early on are at least partially convinced about their
applicability and usefulness. Therefore it may be relevant for the reader of this
book to reverseor at least consider doing sothe order in which the chapters are
read or to split up the reading into appropriate parts. Therefore if you as reader is
curious to find out whether the book offers some interesting news, an option may
be to read Chap. 7 on the case example first and then Chap. 8 which summarises
the main findings and recommendations. Then you may continue with Chaps. 5
and 6 with their presentations of tools and methodology. Chapters 24 about the
v
vi
Preface
grounding of the later practice are probably the most heavy-going and may
therefore as suggested above be postponed and skipped in the first place. Chapter 1
is an introduction that, among other things, sets out the main themes treated in the
book and gives an overview of the chapters. To facilitate the reading of the book
and tie the individual chapters together, each chapter ends with a listing of the
main points and findings of that chapter. Additional technical material is presented
as two appendices.
The SP approach treated in the book has so far been tested on a number of cases
which has shown its potential to those involved. I sincerely hope that Complex
Strategic Choices will inspire and support readers in their dealing with upcoming
complex planning problems and strategic decision making.
Virum, Denmark, October 2011
Steen Leleur
Acknowledgments
In connection with the completion of the book, I would like to thank the following
for collaboration about systemic planning (SP) cases and the implementation of
methodology: Assistant professor Kim Bang Salling, research assistant,
Ph.D. student Anders Vestergaard Jensen and research assistant, Ph.D. student
Michael Bruhn Barfod. I also thank the many students that have taken my planning
and appraisal courses at DTU Transport in the last five years for their useful
feedback on the SP theory and methodology. My thanks also go to the many
participants in Denmark and Sweden that over the last five years have attended SP
decision conferences and contributed in a very constructive and positive way. This
has been of great value to the formulation and adjustment of the practice framework described later in the book. A successful decision conference is in the heart
of systemic decision support in accordance with the SP principles.
Thanks also to professor Michael C. Jackson, professor Frank Stowell and
professor Cathal Brugha who all provided creative suggestions to an early outline
of this book. Furthermore, I am grateful for constructive comments to a draft
version of the book provided by professor Jotin Khisty and professor Alan
Pearman and to both of them also for enjoyable research collaboration over many
years. My writing in the final stage gained from my attendance as a teacher at the
Helsinki Summer School in Transportation 2011. I owe thanks to professor Antti
Talvitie for inviting me and letting me present some of the ideas contained in this
book and for his valuable comments. I also want to thank DTU Transport director
Niels Buus Kristensen for offering me a month free of duties to make it possible
for me to concentrate on the manuscript. Furthermore, I appreciate research grants
from the Danish Strategic Research Council and the EU Regional Development
Fund, Interreg IV-A, for my work in the research projects UNITE (20092012)
and EcoMobility (20102012), which have benefited the development of the ideas
presented here.
Senior editor Anthony Doyle at Springer deserves thanks for having originally
invited me to write a book in Springers Decision Engineering Series. He and his
assistants Claire Protherough and Grace Quinn at Springer furthermore deserve
my thanks for their kind help and support in the publishing process. Also thanks
vii
viii
Acknowledgments
Contents
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
1
1
2
2
3
5
5
7
9
12
13
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
15
15
16
17
18
19
20
21
25
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
27
27
28
29
30
30
31
31
33
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
ix
Contents
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
35
35
37
38
39
41
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
43
43
48
49
50
53
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
55
56
57
58
59
62
62
63
64
65
66
67
68
69
..
71
..
..
71
72
..
..
..
73
76
76
..
78
..
82
..
85
..
85
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
Contents
xi
7.2.6
95
95
95
97
108
109
112
112
115
116
119
Appendix A: COSIMA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
123
Appendix B: SIMDEC. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
149
Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
165
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
90
92
93
Abbreviations
AHP
AI
BCR
BS
CAD
CBA
CDM
CEA
CEO
CF
CG
CGE
CLG
COSIMA
CPM
CSH
CSR
CST
CV
DA
DE
DM
DSS
DTU
EIA
FA
FRA
FW
FYB
GIS
xiii
xiv
IP
IRR
ITS
MCA
MCS
MM
MOE
MP
NPV
OR
PA
RA
RCF
REMBRANDT
ROC
ROD
RSF
SA
SC
SD
SIMDEC
SMART
SMARTER
SP
SROI
SSM
STA
SW
SWOT
TGB
TRR
TSI
TV
UU
VF
VTS
Abbreviations
Interactive planning
Internal rate of return
Intelligent traffic service
Multi-criteria analysis
Monte Carlo simulation
Mind mapping
Mode of enquiry
Multiple perspectives
Net present value
Operations research
Preference analysis
Risk analysis
Reference class forecasting
Ratio estimations in magnitudes
or deci-bells to rate alternatives
which are non-dominated
Rank order centroid
Rank order distribution
Reference scenario forecasting
Scenario analysis
Strategic cognition
Systems dynamics
Simulation and multi-criteria analysis
for decision making
Simple multi-attribute rating technique
Simple multi-attribute rating technique
exploiting ranks
Systemic planning
Social return on investment
Soft systems methodology
Stakeholder analysis
Swing weights
Strengths, weaknesses, opportunities
and threats
Traffic Plan GreenlandDecision Tool
(transl. from Danish)
Total rate of return
Total systems intervention
Total value
Unknown unknowns
Value function
Vessel traffic service
Chapter 1
The four levels are treated as follows: Level 1 of the SP framework is treated
in Chaps. 2 and 3, level 2 in Chap. 4, level 3 in Chaps. 5 and 6 and level 4 in
Chap. 7. This chapter and Chap. 8 respectively serve to introduce the topic of
complex strategic choices and to summarise and validate its findings together
with some advice on practice for readers who may wish to make use of SP for
addressing complex strategic choices. The contents of the chapters are overviewed below.
particular epistemic lens. In systemic planning this Complexity paradigm functions as a complement to the Simplicity paradigm.
Chapter 4 on The systemic process introduces the American brothers Hubert
and Stuart Dreyfus by treating their theory of learning. With a background in
philosophy and in operations research (OR) respectively, they have formulated a
learning process into five steps, which deals with what they themselves
acknowledge as representing a development from novice to expert. Although
apparently not familiar with Morins Simplicity and Complexity paradigms, they
describe a learning cycle which in its progression can almost be seen to build on
the paradigm of simplification in combination with the paradigm of complexity. In
this context it leads to the idea of designing a kind of systemic, self-organising
learning process to be made use of in the set-up of systemic planning.
With the theory outline of systemic planning presented in the above chapter,
Chap. 5 on The systemic toolbox addresses the more methodical and practical
aspects of SP. Specifically a number of useful OR methods and techniques are
introduced as a kind of long list for methods of possible relevance for SP. These
are categorised and presented in the form of three waves, with each new wave
we are still in the third wave since the early 1990srepresenting an important new
approach direction. From the long list of OR methods seven hard and seven soft
methods are identified, which are seen to provide a suitable method arsenal in the
subsequent practical adaptation of systemic planning.
In the following Chap. 6 on Setting up the decision support focus is set upon
applying different modes of enquiry based on the findings in the previous chapter.
Next the scoping of strategic choices is addressed and afterwards the assessment of
consequences and risks is treated. Emphasis is placed on describing the interrelatedness of scoping and assessment in the process of establishing adequate
decision support for complex strategic choices. Based on the specific purposes of
scoping and assessment a number of both hard and soft methods are treated. This
chapter ends with an outline of what can be seen to characterise a suggested
concept of choice intelligence.
The following Chap. 7 on Company relocation as demo-case serves to
exemplify systemic planning in the form of a description of a complex planning
task concerning the relocation of TRANS-IT Consult. The case concerns the
application of SP for selecting a new company headquarters location in the
resund region with many different factors influencing the final choice among a
set of eight pre-screened possibilities. The case description includes the various SP
steps and considerations leading towards a final strategic decision about the most
attractive new location.
The concluding Chap. 8 A summing up: The challenge of strategic decision
making first reiterates some of the main concepts of SP and afterwards ten cases
where SP has been applied are reviewed. This forms the background for a subsequent assessment of the validity and potential of the SP framework. Afterwards
complex strategic choices are put into a wider context, where issues about known
and unknown and risk-related Black Swan theory are made use of to indicate what
types of challenges organisations and companies may face with regard to long-term
planning and complex strategic decision making. Finally some conclusions are
presented together with a developmental perspective on SP.
The book also contains two appendices giving a more technical description on
how cost-benefit analysis (CBA) and multi-criteria analysis (MCA) can be combined in the COSIMA methodology in Appendix A and multi-criteria analysis and
risk analysis (RA) in the SIMDEC methodology in Appendix B. The CBA, MCA
and RA are all included in the systemic toolbox described earlier in Chap. 5.
Detail complexity helps us focus on the influences from the system demarcations and the system components as they enter at an early stage in our examinations and/or models. Seemingly, the system is something that is given at the
beginning of a study. This view, however, is much too simple because beneath its
mere representation the system is also the result of a history that has frozen into
the concrete system elements and their interrelations. To demarcate the system
properly, we need to become aware of the details and their possible meaning and
influence. This kind of awareness is made explicit to us through the work of
physicists on complex systems. One important finding is that so-called critical
states are ubiquitous. In socio-technical systems the occurrence of critical states is
often what makes problems wicked.
The need to pay attention to the details is well argued in the quotation below
from Mark Buchanan, a theoretical physicist now working as a science writer:
By studying the natural kinds of patterns that evolve in networks of interacting things
under non-equilibrium conditions, we may be able to understand an immense range of
natural phenomena, from our turbulent atmosphere to the human brain. The study of
complex systems is all about things that are out of equilibrium, and on this task, of course,
scientists are really just starting out. So the relationship between the critical state and
complexity is really quite simple: the ubiquity of the critical state may well be considered
the first really solid discovery of complexity theory.
And yet there is another useful way to look at all this. In coming to consider complex
systems, physicists seem to have gained a new appreciation of a simple fact: in the
immediate world around us, history is important. For living things, which ultimately
develop from a single cell, this is obvious. But one cannot even understand the hardness of
a steel pipe, or the irregular surface of a fractured brick, without referring to the full
history of its making (Buchanan 2001, p. 16).
For these reasons boundary setting becomes a major influence when defining an
adequate action area to identify, for example, the possible means to secure
improvement. At this stage when presenting the theory behind systemic planning
to be applied for providing decision support, it suffices to say that the demarcation
of the socio-technical system that we are examining is by its nature less given
than first impressions might suggest.
The medium in which detail complexity operates is typically space (covering
resources such as persons and their skills, physical facilities, financial resources,
etc., which make up the variables in this space). Clearly with many variablesand
with each variable possibly having many attributes of relevance and with interdependence between variablesthe detail complexity becomes full-fledged.
Briefly stated, detail complexity relates to concerns about means.
Fig. 1.2 Weather sequences in a computer model: the Butterfly Effect by Lorenz. In 1961 the
meteorologist Edward Lorenz found that small differences in starting conditions could mean a
considerable change in end result. Thus a storm at one location may be seen as initiated by a
butterfly flapping its wings and thereby causing a small disturbance up-stream of the weather
pattern propagation that resulted in the storm. Adapted from Gleick (1987, p. 17)
give less attention to detail complexity than to dynamic complexity (Senge 1990;
Flood 1999, pp. 1314).
In Fig. 1.2 the importance of dynamic complexity is illustrated by comparing
the development of two weather patterns.
With nearly the same starting point, the two patterns diverge over time and end
up with no resemblance at all.
The work of Edward Lorenz in the 1960s was very important in initiating
research on chaos in dynamical systems, although deterministic chaos as a
phenomenon had been known for many years due to the work of, among others,
the French mathematician Henri Poincar around 1900. The use of computers
has come to play a major role in the research that started with the findings of
Lorenz.
With the focus on planning and decision making, we have to interpret the
importance of dynamic complexity by the way it makes long-term forecasting a
highly doubtful undertaking. But many further insights are implied when we
examine complex dynamic interrelationsnot least if our focus is more on human
organisations and their development than on weather pattern propagation. Perhaps
in this organisations context we should see the Butterfly Effect as a storm
started, for example, by the whispering of a rumour one afternoon at the coffee
machine.
The organisations and chaos researcher, Ralph Stacey, has given the following
interesting interpretation of organisational time dynamics (speaking in the context
of the phenomenon of change) by making reference to the so-called leverage
points introduced into systems vocabulary by Peter Senge. With a focus on
studying business units as complex, dynamic systems Stacey says:
The study of complex, dynamic systems provides the insight that the behaviour of a system
cannot be understood simply by examining the systems parts. The system in effect has a
life of its own. The system itself has a major impact on behaviour and therefore on
outcomes. Thinking therefore has to proceed in terms of whole systems, their interconnections, and the patterns of behaviour they may generate. Changes accumulate slowly out
of the interconnections between a systems parts. Focusing on snapshots of the parts,
looking for cause-and-effect links that are close together in time and space, means missing
the slow accumulation of change. Instead of trying to understand quantitative detail of
parts, therefore it is far more fruitful to try to understand the qualitative nature of
interconnections and patterns of behaviour. It is especially helpful to try to find the points
in the system that are most sensitive and amplifyingthe points of greatest leverage. By
operating at these points rather than trying to control details everywhere, managers can
bring about the greatest changes in the system with the least effort (Stacey 1993, p. 110).
Peter Senge and his collaborators have managed to identify a number of what
he calls archetypes of change, which are dynamic organisational patterns. One
of these is the Tragedy of the Commons, which occurs when two systems
operate in the same environment and are rewarded initially by exploiting the
environment (Jackson 2000). The tragedy of the commons was originally coined
by Hardin (1968) in an article in the journal Science, in which he examined
individual actions and their cumulative consequences which, in an unwitting way,
could be systematically destructive for the socio-economic unit made up of the
individual actors. His picture was the medieval English village where each
householder made the apparently reasonable decision to graze as many cattle on
the commons as possible with the result that the commons would suffer overgrazing, leaving each and every householder in a poorer condition.
There is no doubt that a number of archetypesPeter Senge operates with
around a dozencommunicate what we would like to see as collectively gained
lessons that are of importance with regard to interpreting possible development
patterns. Clearly, they play a major role for the manager who does not want to
embark on some kind of course that may later turn out to be less desirable for some
reason. However, downplaying the checking of details cannot generally be
recommended, see the quotation above, nor can the belief that a relatively limited
number of archetypes is capable of unfolding a larger part of the dynamic complexity relating to change in business units or to change in socio-technical systems
in general. As with detail complexity and its means-uncertainty when defining the
action area around complex strategic choices, dynamic complexity indicates
another major type of uncertainty as a basic condition.
The medium in which dynamic complexity operates is time and stated briefly
dynamic complexity relates to concerns about path.
10
Simon via his writings about organisational decision making (1968) and the
sciences of the artificial (1969), the first major exposition of the meaning and
consequences of applying the view of organised complexity on organisations as
systems. Another major influence is Jrgen Habermas (1979, 1986, 1989) with his
elaborate theory on communication. The following quotation by McCarthy from
his magnificent book The Critical Theory of Jrgen Habermas brings the third
type of complexity to the fore, namely what in this presentation is termed the
complexity of interests (following the German term) or simply preference
complexity. Thus McCarthy with reference to Habermas states that:
a precondition of rational consensus is the thematisation of available need interpretations themselves; interests are neither empirically found nor simply positedthey are
shaped and discovered in processes of communication (McCarthy 1981, p. 328).
What the theory of Habermas states here is that preferences (interests) are
tied up with processes of communication and are therefore quite dependent on the
issues raised and debated. To deal with the complexity involved and get to grips
with the interests that might be associated with the various stakeholders, we need
to understand the processes of communication. Normally when referring to theory
of communication we have the work of the mathematician Claude Shannon in
mind. Contrasting, however, Shannons theory of communication with the version
stemming from Habermas we find that the two theories of communication are
completely different. Making use of the complexity notions I have introduced, we
might say that Shannons theory deals with a measurement of message transmission (with the complexity issues involved then relating more to the notion of
detail complexity), whereas Habermass theory examines the basic components of
human language and interaction (based on what he calls validity claims). For our
purpose the theory of Habermas gives the valuable insight that preferences are not
ready-made and accessible for strategic decision making but have to be shaped
and discovered. As can be seen later on this insight has a very practical imprint
on the way that systemic planning is carried out, namely as a kind of search-learndebate process.
The notion of preference complexity can be illustrated as shown in Fig. 1.3
which indicates the shaping of an interest in a symbolic way.
The figure illustratesin a symbolic way onlythe complexity involved when
shaping and defining an interest. In part (a) we have all possible fragments and
influences which in part (b) have obtained a certain degree of common orientation.
What the figure really shows is atomic magnets in a piece of iron: at high temperatures (a) they cannot line themselves up due to thermal jostling, but at lower
temperatures (b) they are able to align with the result that the iron becomes
magnetic. For the purpose here as mentioned the figure is only symbolic and
illustrative. Thus we can perceive the figure as showing a heated debate that may
(or may not) be cooling off and lead to clarification and explication of a certain
interest, depicted as a change from (a) to (b).
Preference complexityand the set of related issues to be addressedhas not
had its proper role in the development of concepts and tools in systems science
11
(Leleur 2008). This may be due to the main professions involved in its development over the five-to-six decades since the Second World War, with scientists,
engineers and economists dominating with regard to theory and practice on the
basis of the terms and premises of their educational background. However, similar
to detail and dynamic complexity when defining an adequate action area around
complex strategic choices, preference complexity is an important issue.
The medium of preference complexity is mind and briefly stated, preference
complexity relates to concerns about ends.
In this way we have obtained complexities that operate in space (detail complexity), in time (dynamic complexity), and in mind (preference complexity).
Later onafter dealing with the basic theories behind the complexitieswe will be
able to see that exploring complex strategic choices relates clearly to all three concerns: means, path and ends. A main theme of the book is the attention decisionmakers in general will have to pay to complexity issues. I agree with Senge in
believing in the importance of dynamic complexity, but pay at least as much attention
to detail complexity. I pay attention to preference complexity, because I have come to
believe that insights into this type of complexity have a special role to play when
addressing strategic decision making. The theories and methods that are presented in
the following chapters all relate to the impact that complexity has on the way decision
support can be provided.
In the following section I end this introductory chapter by addressing a general
classification of problem types of relevance for planners and decision-makers. In
the final chapter of the book this classification will be reconsidered to take account
of the ways in which the systemic planning approach set out can be seen to add
value to current knowledge about ways of qualifying the process and methodology
relating to making complex strategic choices.
12
Table 1.1 Problem types relating to the configuration of means and ends
Problem types
Four different configurations and related approaches
Means/ends
Certain
Uncertain
Certain
A: Computation
B: Judgement
Uncertain
C: Compromise
D: Chaos or Inspiration
13
References
Buchanan M (2001) Ubiquitythe science of history or why the world is simpler than we think.
Phoenix
Flood RL (1999) Rethinking the fifth disciplinelearning within the unknowable. Routledge,
London
Gleick J (1987) Chaosmaking a new science. Viking, New York
Habermas J (1979) Communication and the evolution of society. Heinemann, London
Habermas J (1986) The theory of communicative action, Volume 1: Reason and the
rationalization of society. Polity Press, Cambridge
Habermas J (1989). The theory of communicative action, Volume 2: The critique of functionalist
reason. Polity Press, Cambridge
Hardin G (1968) The tragedy of the commons. Science 162:12431248
Jackson MC (2000) Systems approaches to management. Kluwer Academic, New York
14
Chapter 2
15
16
Social systems which are subdivided into: societies, organisations and interactions
Psychic systems
Organisms and
Machines
For the upper three types of systems, Luhmann sees the autopoietic perception
as relevant, whereas machines are categorised as allopoietic, i.e. their different
parts or units are not maintained and produced by themselves during the operation
of the machinethey are excluded as part of the machine system in such a way
that they can be replaced if, for example, a certain unit is worn out. This is why the
originators of autopoietic systems thinking, the Chilean biologists Varela and
Maturana, see only living systems as autopoietic.
Social and psychic systems are prominent as concerns the attention they are
given in Luhmanns work. Basically, these systems are characterised by their use
of meaning, something which is not attributed to the categories of organisms and
machines. For our purpose I will concentrate on the social systems category, but
the perception of a system being operationally closed can be illustrated in the
psychic systems category: the nervous system of the brain is a self-referential,
closed system consisting of neurons. An event does not directly determine neural
activity, but may have an effect that can be seen as a kind of irritation. The closed
nature of the system is then not to be understood as a kind of preclusion but rather
as a possibility for openness. The neural system does not mirror the surroundings
but constructs an image of these through internal operations.
17
18
19
these manifestations could have been otherwise and have led to other configurations with other interpretations, or with other understandings of these interpretations; other perspectives could have been chosen and in that case this or that
consideration or reflection would change. When we supplement the view of the
decentred approach to systems thinking with the views expressed above on the
notion of contingency, we arrive at some of the basic analytical tools in
Luhmanns treatment of social systems.
20
system to be on alert all the time. Being on alert on this basis and acting in
accordance is what management is about. SP, at least in those contexts where this
type of approach may be applied, can be seen as one of the tools available for
management, but an important one, as it is precisely aimed at assisting the management and the organisation in facilitating what Luhmann calls the temporalization of complexity.
Systemic planning, therefore, has a role that in some ways is different from
conventional planning: while the latter aims at controlling the march into the
futureseen as part of the present systems environmentSP accepts that this
environment is to some extent uncontrollable. More practice-oriented, it makes use
of a framework and methodology that make the organisation better prepared to
meet the future by recognising both its knowns and its unknowns.
21
22
The latter statement about not knowing the consequences in advance of actions
to be taken is really one of Staceys strong points. His considerations have led him
to speak about the unknowable. His viewpoint is expressed in the quotation
below:
Everyone admits that the future is basically unknowable, particularly in the case of an
innovative product or course of action. This prospect, however, makes many managers
uncomfortable, and they then ease their discomfort by assuming that even innovative
futures are nonetheless approximately knowable. One can at least, they say, have a vision
or make some assumptions about the long-term future. One can give shareholders, or
others in a controlling position, meaningful information on future rates of return and risk
levels.
I argue that this is a soothing fantasy that distracts attention from, and weakens the
resolve to deal with, the real world. Instead of sidestepping the issue of unknowability,
managers must learn to face it head on. That means accepting that you really have no idea
what the long-term future holds for your organisation; forming visions and making
assumptions are not realistic possibilities. It means accepting that no individual or small
group can be in control of an organisations long-term future (Stacey 1993b, p. 7).
One can agree with many of Staceys findings in his comprehensive writings
about organisations and issues relating to change (Stacey 1993a, b; Stacey et al.
2000). However, I cannot agree with his very principal meta-finding quoted above
that one should recognise and accept that no individual or small group can be in
control of an organisations long-term future. On the contrary, the viewpoint
argued in what follows is that proactive effortseen as planning in its broadest
terms and handling complex strategic choices properlyis worthwhile. What is
basically meant by this is that it is necessary to scrutinise the consequences of an
action in advance of any concrete action when the consequences can be identified
and assessed. This is exactly the job of managers assisted by planners and decision
analysts.
Conventionally such a mapping of consequences has relied in a comprehensive
way on modelling and quantitative assessment, in this book referred to as systematic planning. This conventionalsystematicperception of planning implies
causality: if these means are used in a specific way, certain ends will result from
or be caused bythe set of planned actions. Such if then thinking belongs to the
generic idea of planning and strategy development.
However, since chaos and complexity theories entered management thinking in
the 1990s, the concept of causality in socio-technical systems has been under
attack. Above I used the work of Ralph Stacey to exemplify what may come out of
basing management thinking on these new theory constructs. I could have chosen
several authors of management literature, but Staceys books about chaos management published in the early 1990s stand out due to their penetrating insights
about organisations and not least their in some respects radical conclusions (Stacey
1993a, b). For our purpose here, seeking to come to terms with planning under
complex conditions, I highlighted the finding from this type of literature that no
individual or small group can be in control of an organisations long-term future.
This is certainly not the common belief among board members and CEOs and
23
therefore there is a need to dig deeper into this question. Summarising in Chap. 8
some plausible answers will be presented based on the findings throughout the pages.
With a focus now on causality, I will make use of Staceys more recent work
about complexity and management (Stacey et al. 2000). Due to debates with
several well-known researchers in the field, among others Jonathan Rosenhead,
Staceys argumentation on the implications of chaos and complexity theories for
management has somewhat changed. Stacey and his collaborators now recognise
that a deterministic, chaotic system may not resemble the ongoing affairs of human
organisations where many uncertainties could better be comprehended as being of
a stochastic type. The early writings of Staceythose used to introduce the
concept of dynamic complexity in Sect. 1.3focus some attention on the idea of
the unknowable future to be accepted on the premises chosen; we may see the
more recent book from 2000 by Stacey, Griffin and Shaw as a continued examination of the unknowable which is, among other things, carried out by digging
deep into the concept of causality. In fact the book, entitled Complexity and
Management, contains a thorough description and interpretation of causality in
Western thinking from Aristotle to the present day, in which, for example, the
theories of self-organisation worked out by the Belgian physicist Ilya Prigogine
and others are dealt with.
The major finding of Stacey and his collaborators centres around the concept of
transformative teleology. This is contrasted with the concepts of formative
teleology, which are concerned with the actualisation of form or self already there
in some sense and rationalist teleology concerned with realising chosen, pre-set
goals. Telos is the Greek word for goal or end, so teleology can be understood
as preoccupied with the overarching source of change (Ibid., p. 196). We can get
an understanding of the causal framework Stacey and his collaborators associate
with transformative teleology from the following quotation:
When Prigogine considers the wider implications of his work, we think he makes a clear
move to Transformative Teleology. At the beginning of his book, The End of Certainty
(1997), he poses what he sees as a central question: Is the future given, or is it under
perpetual construction? In the terms we are using, this translates into Is causality in
nature (including humans) better understood as Formative Teleology, or is it better
understood as Transformative Teleology? His answer to the question is very clear: he
sees the future for every level of the universe as under perpetual construction and he
suggests that the process of perpetual construction, at all levels, can be understood in
nonlinear, nonequilibrium terms, where instabilities, or fluctuations, break symmetries,
particularly the symmetry of time. (Stacey et al. 2000, p. 97).
The breaking of time symmetry plays an important role in the more technical
parts of Prigogines work (Prigogine and Stengers 1985, pp. 249341). The
interest pursued by Stacey, Griffin and Shaw, however, is in exploring their
conceptual implications as expressed in the quotation below:
[Prigogine] says that nature is about the creation of unpredictable novelty where the
possible is richer than the real. When he moves from focused models and laboratory
experiments to think about the wider questions of evolution, a move that many scientists
would question, he sees life as an unstable system with an unknowable future in which the
irreversibility of time plays a constitutive role. He sees evolution as developing bifurcation
24
points and taking paths at these points that depend on the micro details of interaction at
those points. Prigogine sees evolution at all levels in terms of instabilities, with humans
and their creativity as part of it. For him, human creativity is essentially the same process
as natures creativity and this is the basis for his call for a new dialogue with nature.
These features, unknowable futures emerging in here-and-now interactions, are essentially
what we have defined as the causal framework of Transformative Teleology.
Central to Prigogines approach, at all levels, is the distinction between individual
entities and populations, or ensembles, consisting of those entities. (Stacey et al. 2000,
p. 97).
When the self-organisation phenomena are interpreted later in the text with
emphasis on human organisations, the following conclusion is reached:
the dominant management discourse, including systems thinking, is built, explicitly or
implicitly, on Rationalist Teleology as an explanation of choice This is expressed in
psychological theories that accord priority and primacy to the choosing individual over
the social. It is a view of minds as information processing devices that make representations of a pre-given world, formed into maps and models that are the basis of subsequent
action. Alternatively, individuals may be thought of as having deep, true identities and
they are motivated, ultimately, by contexts that allow them to express their true natures.
The socialthat is, the cooperative and competitive relating between peopleis
important as an enabling context (Stacey et al. 2000, p. 181).
25
References
von Bertalanffy L (1972) General system theorya critical review. In: Buckley W (ed) Modern
systems research for the behavioral scientist. Aldine Publishing Company, Chicago
von Bertalanffy L (1973) General system theoryfoundations, development, applications.
Penguin books, London
Checkland P (1981) Systems thinking, systems practice. Wiley, Chichester
Harste G (1992) Niklas Luhmanns konstruktion af samfundsteori. In: Jacobsen C (ed)
Autopoiesis: En introduktion til Niklas Luhmanns verden af systemer. Forlaget politisk
revy, Copenhagen
Kneer G, Nassehi A (1997) Niklas Luhmannintroduktion til teorien om sociale systemer. Hans
Reitzels Forlag, Copenhagen
Leleur S (2008a) Systems science and complexity: some proposals for future development. Syst
Res Behav Sci 25(1):6779
26
Leleur S (2008b) Systemic planning: principles and methodology for planning in a complex
world, 2nd edn. Polyteknisk Forlag Lyngby, Denmark
Leleur S (2008c) At navigere mod fremtiden: Systemisk planlgning som ide og metode.
Polyteknisk Forlag Lyngby, Denmark
Luhmann N (1995) Social systems. Stanford University Press, Palo Alto
Moeller H-G (2006) Luhmann explained: from souls to systems. Open Court, Chicago
Prigogine I, Stengers I (1985) Den nye pagt mellem mennesket og universet: Nye veje i
naturvidenskaberne. Forlaget ASK, Copenhagen
Stacey RD (1993a) The Chaos Frontiercreative strategic control for business. ButterworthHeinemann Limited, Oxford
Stacey RD (1993b) Managing the unknowable. Josey-Bass Publishers, San Francisco
Stacey RD, Griffin D, Shaw P (2000) Complexity and managementfad or radical challenge to
systems thinking. Routledge, London
Thyssen O (1992) Forhold som forholder sig til sig selv. In: Jacobsen C (ed) Autopoiesis:
En introduktion til Niklas Luhmanns verden af systemer. Forlaget politisk revy, Copenhagen
Thyssen O (1997) Vrdiledelseom organisationer og etik. Gyldendal, Copenhagen
Chapter 3
27
28
29
Simplicity paradigm
Complexity paradigm
Universality
Determinism
Dependence
Necessity
Lawfulness
Prediction
Separation
Identity
The general
Objects
Elements
Matter
Quantity
Linear causality
The automaton
Objectivity
Multiplicity
Organisation
Autonomy
Possibility
Self-organisation
Surprise
Wholeness
Individuality
The particular
Subject
Interactions
Life
Quality
Multi-causality
Time
Culture
30
31
Later on he describes more closely how an interaction between the whole and
the parts can take place:
we can make a higher level of understanding based on the constructive circularity of
the explanation of the whole by the parts and the parts by the whole, in which the two
explanations become complementary in the movement associating them without losing
their simultaneous and opposing characteristics (Morin 1992, p. 131).
3.3.2 Reflection-in-Action
Focusing on a practical application of theoretical insights from Luhmann and
Morin, we can draw on the American organisation theorist Donald Schn, who
developed a reflection-in-action inquiry that deals with learning processes at
individual and group level (Schn 1983). In Chap. 4 we will explore the role
learning processes can play within the systemic approach by making use of the
brothers Hubert and Stuart Dreyfuss five-step learning model.
It is a striking feature that the insights contained in the theory elements from
Luhmann, Morin, Schn, and Dreyfus and Dreyfus in many ways resemble each
other. In this respect Schns work seems to constitute a good and practical bridge
between the general theoretical insights of Luhmann and Morin to the concrete
model of learning by Dreyfus and Dreyfus, who like Schn focus on individuals
learning processes.
32
Problem formulations are necessarily biased by methods, objectives, perceptions, etc. As stated above, we do not expect an engineer and a sociologist to
approach a problem in the same way. Rather schematically, we may expect the
engineers approach to be closer to the Simplification paradigm than the approach
taken by the sociologist and vice versa as concerns the Complexity paradigm. With
a systemic approach we have, however, paved the way for a combination of the
two basic epistemic lenses represented by the paradigms of simplicity and
complexity.
As already noted above, learning takes on a particular role when engaging in
complementing conventional thinkingreferred to as being systematic by naturewith a type of thinking referred to as being systemic by nature. These ways
of thinking about a complex strategic issue are our first level of the systemic
framework introduced as four interconnected levels in Sect. 1.1. By means of the
two basic epistemic lenses of simplicity and complexity we have established ways
of seeing, which will link complexity and simplicity represented later on by
learning in association with quantitative modelling. Chapter 4 will address
learning processes.
Main points and findings of this chapter
According to Edgar Morin the exploration of reality makes two basic epistemic
lenses necessary. By referring to the epistemic lenses as paradigms these are the
Simplicity paradigm and the Complexity paradigm.
The Simplicity paradigm and the Complexity paradigm are not meant to replace
each other but should be made use of in a complementary way.
33
Furthermore Morin states that complexity is not a surface noise of the real, but is
the very principle of the real.
A fundamental principle is set out by Morin as constructive circularity, where
the explanation of the whole by the parts and the parts by the whole constitutes
our cognition of the real.
The considerations about the epistemic lenses lead to the findings that systemic
planning should make balanced use of simplicity and complexity thinking and
adopt constructive circularity as stepping stones in formulating the SP
framework.
References
Capra F (1982) The turning pointscience, society and the rising culture. Flamingo by Fontana
Paperbacks, London
Hansen S, Kolmos A (1998) Projektvejlederen som mesteren i en gensidig forstelsesdialog,
Pdagogisk Udviklingscenter, Aalborg Universitet
Kuhn TS (1962) The structure of scientific revolutions. Chicago University Press, Chicago
Leleur S (2008) Systems science and complexity: some proposals for future development. Syst
Res Behavioral Sci 25(1): 6779
Morin E (1974) Complexity. Int Soc Sci J 26(4): 555582
Morin E (1985) Kompleksitetens bud. Paradigma 1(1):1820
Morin E (1992) The concept of system and the paradigm of complexity. In: Maruyama M (ed)
Context and complexity: cultivating contextual understanding, Springer, Heidelberg,
pp 125138
Schn DA (1983) The reflective practitioner. Basic Books, New York
Chapter 4
35
36
a real-world setting. To cope with such problems, people have to learn to adopt a
hierarchical procedure of decision making to organise the situation. By first
addressing a small set of the most important factors, a persons performance can
both be simplified and improved. As Dreyfus and Dreyfus explain, a competent
performer with a goal in mind sees a situation as a set of facts whose individual
importance may depend on the presence of other facts. Given a situation with a
particular constellation of facts, a certain conclusion should be drawn, a decision
made, or an expectation investigated. In the car-driving example, the safe and
smooth operation of the vehicle is related, for example, to the goal in mind of
going quickly from A to B. The route is then chosen with attention to distance and
traffic, which may result in manoeuvring the car in a way that disregards passenger
comfort.
Important changes take place from the novice stage via advanced beginner to
the competence stage according to Dreyfus and Dreyfus:
the novice and advanced beginner recognize learned components and then apply
learned rules and procedures. As a consequence, they feel little responsibility for the
outcome of their acts. Assuming that they have made no mistakes, an unfortunate outcome
is viewed as the result of inadequately specified elements or rules. The competent performer, on the other hand, after wrestling with the question of the choice of a plan, feels
responsible for, and thus emotionally involved in, the product of his choice. While he both
understands and decides in a detached manner, he finds himself intensely involved in what
occurs thereafter. An outcome that is clearly successful is deeply satisfying and leaves a
vivid memory of the plan chosen and of the situation as seen from the perspective of the
plan. Disasters, likewise, are not easily forgotten (Dreyfus and Dreyfus 1988, p. 26).
37
Context-free
Context-free and
situational
Context-free and
situational
(4) Proficient
Context-free and
situational
Context-free and
situational
(5) Expert
Commitment
None
None
Analytical
Analytical
Detached
Detached
Chosen
Analytical
Experienced
Analytical
Experienced
Intuitive
Detached understanding
and deciding. Involved
in outcome
Involved understanding.
Detached deciding
Involved
38
With regard to rationality and intelligence, this makes room for arationality.
This is argued in the following way (Ibid., p. 36):
The moral of the five-stage model is: there is more to intelligence than calculative
rationality. Although irrational behaviorthat is, behavior contrary to logic or reason
should generally be avoided, it does not follow that behaving rationally should be
regarded as the ultimate goal. A vast area exists between irrational and rational that
might be called arational. The word rational, deriving from the Latin word ratio, meaning
to reckon or calculate, has come to be equivalent to calculative thought and so carries
with it the connotation of combining component parts to obtain a whole; arational
behavior, then, refers to action without conscious analytic decomposition and recombination. Competent performance is rational; proficiency is transitional; experts act arationally (Dreyfus and Dreyfus 1988, p. 35).
39
40
Table 4.2 The SP generic structure as four interrelated modes of exploration and learning
SP generic
Systemic
Systematic
structure
Scanning
Assessment
In the figure different methods are indicated to illustrate some possible method choice that can
assist in the systemic process. Adapted from Leleur (2003, p. 262)
higher ones) are associated with systems techniques such as some soft OR
methods. The methods and techniques will be the main topic in Chap. 5.
Wrapping up a number of findings the application of systemic planning (SP) for
complex strategic choices is based on having recognised that the challenge facing
the strategic decision problem is related to open-ended change. This means we
must abstain from relying solely on an analytical approach. Instead the SP
approach consists of embedding the final decision to be taken in a process of
building knowledge about the concrete decision task within a subworld that
unfolds in the course of events. Therefore we need to set a team that can be
expected to handle this. Generally this team will consist of analysts and decisionmakers. Later on in Chap. 6 we will consider the use of decision conferences,
where stakeholders of different kinds are also involved. The subworld notion as
introduced above sets focus upon the importance of avoiding unintended closures
when the systemic process unfolds.
The validity of applying SP on complex strategic choices by using the outlined
systemic process will be returned to in the final Chap. 8. At this later stage the
discussion will be underpinned by its application on a number of presented cases.
Main points and findings of this chapter
Hubert and Stuart Dreyfus provide a convincing theory about learning as five
consecutive stages leading from (1) novice via (2) advanced beginner and
(3) competent towards being (4) proficient and (5) expert.
A finding based on this theory of learning made use of in systemic planning is
the importance of going from detached understanding to involved understanding. In SP the systemic process has this transformation of understanding as a
major goal.
Another finding for systemic planning from the Dreyfus five-stage model is that
rationality (what has been referred to as systematic thinking) is but one type of
reasonable cognitive behaviour. At stage 3 in the model, rationality gives way to
what is termed arationality (which is not to be confused with irrationality not to
be recommended). Arationality denotes a kind of intelligent, holistic behaviour
that the Dreyfus brothers associate with the most developed stages of their
model. They simply find that experts act arationally.
Another finding concerns the difference between the cognitive notions subworld
and universe. The latter assumes a mechanistic view where smaller systems (a
41
company for example) are linked into the larger system (a market segment for
example) as a kind of component. The view behind the subworld is opposite to
the mechanistic view as it is seen as a local elaboration of the one commonsense
world we all share. In systemic planning the subworld notion has been adopted
to indicate the collective knowledge about a complex planning problem that
evolves in an SP decision conference. This is exemplified by the description in
Chap. 7 where a subword unfolds around the activities to look for the best
relocation site for a company headquarters. The modes of enquiry to be filled
into the systemic process set out in this chapter are dealt with in the following
Chap. 5.
References
Dreyfus HL, Dreyfus SE (1988) Mind over machine. The Free Press Macmillan, New York
Leleur S (2003) SCOPEan integrated framework for multi-attribute decision making.
Innovation, Carfax publishing 17(3):259270
Leleur S (2004) Systemic decision support in a complex business environment. In: Yucesan E,
Chick S (eds) Proceedings of the 1st future business technology conference (FUBUTEC) at
INSEAD, Fontainebleau published by EUROSIS, March 2004
Leleur S (2007) Systemic planning: dealing with complexity by a wider approach to planning.
Emerg Complex Organ, 98(12):210
Chapter 5
43
44
On the platform of the functionalist and interpretive approaches Jackson and his
collaborators in the 1990s developed critical systems thinking (CST) with
improvement of managerial thinking in mind. They attached CST to Habermass
three cognitive interests: the technical, practical and emancipatory (Habermas
1979; McCarthy 1981). In combination with critical systems heuristics (CSH)
developed by Ulrich (1983), CST constituted a new framework that made it
possible to deal with strengths and weaknesses of the available approaches and the
relations between them, which contributed to emancipatory systems thinking
(Jackson 2000). A great merit is that the work by Habermas hereby was made
accessible and methodologically operational, for example with the total systems
intervention (TSI) methodology from the early 1990s, for professionals involved in
organisational problem solving and decision making. With focus on diversity and
heterogeneity and their meaning and value for managerial decision making,
a postmodern approach has also been formulated which draws on theories by
Lyotard and Foucault (Ibid.; Leleur 2008b).
To support the formulation of the SP framework I have included Tables 5.1 and 5.2
with systems science as three waves on the basis of Jackson (2000, p. 42). The
table was initially set up by Burrell and Morgan in 1979 and later it was extended
by Alvesson and Deetz in 1995 with focus on postmodernism (Ibid., p. 41).
In 2005 I added a fifth column with the purpose of incorporating complexity issues
as part of the third wave (Leleur 2007). On this basiscategorised as three
development waves of systems sciencethe following five modes of enquiry can
be presented in an overview by adequate keywords for a set of features: the
functionalist, the interpretive, the emancipatory, the postmodern and the complexity research orientation. Specifically, the latter is presented in accordance with
the features of a complexity research orientation, which is based on my work with
Systemic Planning (Ibid.). The main sources behind this stem from the previously
described works by Morin and Luhmann. The four other approaches shown in
Tables 5.1 and 5.2 have not been changed in any way as concerns their formulation. In this way Tables 5.1 and 5.2 overview the contents of the first and second
waves of systems science together with ideas which are seen as a tentative profile
of a third wave of systems science that is still unfolding.
In his highly engaging presentation Systems ThinkingCreative Holism for
Managers from 2003 Jackson gives surveys of the functionalist, interpretive,
emancipatory and postmodern research orientationsor paradigms as he prefers to
address them. Over the next pages these surveys are quoted as a background to a
survey of the suggested complexity research orientationor complexity paradigm
to apply this term in the context of paradigm descriptions. Tables 5.1 and 5.2
represent a kind of cognitive billboard, where the patches are the individual
45
Table 5.1 Characteristics of functionalist and interpretive modes of enquiry (MOEs) representing the first and second waves of systems science
Features defining the modes Functionalist mode of enquiry
Interpretive mode of
of enquiry
enquiry
Basic goal
Method
Hope
Problems addressed
Narrative style
Time identity
Organisational benefits
Scientific/technical, strategic
Modern
Control, expertise
Mood
Social fear
Optimistic
Disorder
Organisation metaphor
messages that specify a paradigm. Some of the patches have been found especially
interesting for the work on SP and have been highlighted in the quotations below.
For the functionalist paradigm, which represents the first wave in Table 5.1,
Jackson gives the following description:
The functionalist paradigm takes its name from the fact that it wants to ensure that
everything in the system is functioning well so as to promote efficiency, adaptation and
survival. It is optimistic that an understanding can be gained of how systems work by using
scientific methods and techniques to probe the nature of the parts of the system, the
interrelationships between them and the relationship between the system and its environment. The expertise it provides should put managers more in control of their operations and organizations, and enable them to eliminate inefficiency and disorder.
Associated with this paradigm can usually be found the machine, organism, brain, and flux
and transformation metaphors. (Jackson 2003, p. 38 with accentuation added).
For the interpretive paradigm, which in this context is seen as the second wave
of the development schema for systems science, Jackson has the following
description:
The interpretive paradigm takes its name from the fact that it believes social systems,
such as organizations, result from the purposes people have and that these, in turn, stem
from the interpretations they make of the situations in which they find themselves.
Organizations happen, and people act and interact in organizations, as a result of their
interpretations. This paradigm wants to understand the different meanings people bring
to collaborative activity and to discover where these meanings overlap, and so give birth
to shared, purposeful activity. Managers can be guided to seek an appropriate level of
46
Reclaim conflict
Explore unknown territory
Deconstruction, genealogy Integrate complexity and simplicity
thinking
Claim a space for lost
Contingent insights that will mean
voices
a difference
Carnival
The panopticon with a restricted
view
Marginalisation, conflict Open-ended, wicked and
suppression
hypercomplex problems
Ironic, ambivalent
Multi-dimensional, eclectic
Postmodern
Hypermodern
Diversity, creativity
Awareness, alertness
Playful
Totalisation,
normalisation
Curious
Constrained reasoning and living
47
The quotation below presents a survey of the fifth column about the complexity
paradigm and its features. Similar to the above four paradigm descriptions I have
indicated a metaphor for this paradigm:
The complexity paradigm takes its name from the fact that it believes that many insights
of importance (that make a difference) are contingent in nature due to the complexity of
a messy problem space. It seeks to explore unknown territory in the organizational
landscape with the hope of gaining such insights. The method orientation it makes use of
consists of integrating what it terms complexity thinking with simplicity thinking, where
the latter in many ways resembles the functionalist paradigm. Complexity thinking consists of getting a higher-level understanding based on adopting a constructive circularity
of the explanation of the whole by the parts and the parts by the whole, in which the
complementary nature of this process plays an important role. A demonstrated practice of
such a principle used in planning and referred to as systemic planning (SP) sets focus
upon a self-organising exploration and learning cycle, where alternating systemic perceptions gained as insights by the participants are fed by a combined use of soft and hard
operations research (OR) methods. A toolbox of especially relevant methods and techniques has been worked out consisting of seven soft and seven hard OR methods. The
organisation metaphor is the panopticon with a restricted view. The person in the panopticon gets his viewing capability geared by mirrors and the intricate way that these
can be placed to enable still changing perspectives and surprises. This greater viewing
capability, however, is also tied to an unfolding discovery of hidden corners and sometimes vast areas of dark space that come around at the same time. Thus we have to
recognise that a gearing of knowledge-gaining concurrently implies becoming conscious
of new issues that we will also need to address. Benefits of the complexity paradigm are
then not to be associated with obtaining control or having fun but with benefits due to
awareness and alertness. The mood is therefore not necessarily optimistic, friendly,
suspicious or playfulit might, however, be sobut is basically seen as curious. This
orientation of mood is in good agreement with perceiving the social fear feature in
Table 5.2 as relating to constrained reasoning and living. (Leleur 2009, p. 52 with
accentuation added).
48
wave. In the current third wave, systems as phenomena are no longer clearly out
there or in the mind. As such, one could get the impression that the system
concept per se has become more opaque with regard to explanatory power. This,
however, is not the casein fact the opposite view may be expressed with the
statement that previous understanding of the system notion can be seen as contained in the perception of seeing systems as representations of differences in
complexity (Leleur 2008b).
49
Mode of enquiry
Functionalist
Interpretive
Emancipatory
Postmodern
Complexity
impression is actualised may be both smooth and more disruptive. The panopticon with a restricted view in Table 5.2 may illustrate the occurrence of the
latter: sometimes even a very small change in the angle and shape of a mirror,
symbolising maybe an actual change of methodology, may cause considerably
changed reflections and images due to intricate linkages. The changeto remain
within the metaphoris due to the actual architecture of the allegoric, subworldencompassing mega hall of mirrors and relates to the actual planning conditions.
The latter can be perceived as the problem-specific impacts stemming from at least
detail, dynamic and preference complexity. The images are what we refer to as the
systemic perceptions.
50
51
Soft methods
It is characteristic of the two columns in Table 5.5 indicating the hard and the
soft methods, respectively, that there is a general movement towards more
demanding methods and techniques when reading down through the columns.
One fundamental question is how many hard and soft methods, respectively,
must be involved to ensure a smooth running of a systemic process. There is no
clear-cut answer here.
In the simplest cases, a combination of cost-benefit analysis (CBA) and brainstorming (BS) may lead to a conviction among the decision makers involved that
the factors found by brainstorming and afterwards used to inform a more or less
standardised CBA have provided a sufficient basis for their decision making.
52
Table 5.5 Seven hard and seven soft methods and techniques
The systemic toolbox sustaining the five MOEs
Hard methods
Soft methods
Brainstorming (BS)
Mind mapping (MM)
Strengths, weaknesses, opportunities
and threats (SWOT)
Critical systems heuristics (CSH)
Soft systems methodology (SSM)
Stakeholder analysis (STA)
Futures workshop (FW)
53
The five modes have been scrutinised with regard to their potential to inform
and assist the group learning taking place in the SP decision conference. This
has paved the way for two SP components: one concerns the cognitive billboard
and the other the systemic toolbox, see below.
By posing the five paradigms behind the five MOEs together, a kind of billboard
appears with a multitude of patches. Each patch conveys a message/viewpoint
which may be a discussion trigger in an ongoing group process. In this way the
billboard can contribute to a wide range of deliberations, which of course should
relate to the problem dealt with.
The systemic toolbox consists of seven hard and seven soft operations research
(OR) methods. The methods have been selected from a long list of OR methods
with the purpose in mind that they in various ways should be able to empower
the five MOEs. For the participants in a decision conference this means that a
suitable range of methods and techniques are available which can assist the
group in its deliberations.
References
Churchman CW (1979) Design of enquiring systems. Basic Books, New York
Habermas J (1979) Communication and the evolution of society. Heinemann
Habermas J (1986) The theory of communicative action, volume one: reason and the
rationalization of society. Polity Press, Cambridge
Habermas J (1989) The theory of communicative action, volume two: the critique of functionalist
reason. Polity Press, Cambridge
Jackson MC (2000) Systems approaches to management. Kluwer Academic, New York
Jackson MC (2003) Systems thinking: creative holism for managers. Wiley, New York
Khisty CJ, Mohammadi J (2001) Fundamentals of systems engineering with economics
probability and statistics. Prentice-Hall, New Jersey
Leleur S (2000) Road infrastructure planninga decision-oriented approach, 2nd edn.
Polyteknisk Forlag Lyngby, Denmark
Leleur S (2007) Systemic planning: dealing with complexity by a wider approach to planning,
Emergence: Complex. Org. 98(12):210
Leleur S (2008a) Systems science and complexity: some proposals for future development. Syst
Res Behav Sci 25(1):6779
Leleur S (2008b) Systemic planning: principles and methodology for planning in a complex
world, 2nd edn. Polyteknisk Forlag Lyngby, Denmark
Leleur S (2009) The meaning of system: from CAS to CHESS. The systemist 33(23):4762
(United Kingdom Systems Society)
Linstone HA, Mitroff II (1994) The challenge of 21st century. State University of New York
Press, New York
McCarthy T (1981) The critical theory of Jrgen Habermas. The MIT Press, Cambridge
Midgley G (2000) Systemic intervention: philosophy methodology and practice. Kluwer
Academic, New York
Mingers J, Gill A (1997) Multimethodology. Wiley, New York
Rosenhead J, Mingers J (2001) Rational analysis for a problematic world revisited. Wiley,
New York
54
Troncale LR (1985) The future of general systems research: obstacles potentials case studies
systems research 2(1):4384, Pergamon, London
Ulrich W (1983) Critical heuristics of social planning. Wiley, New York
Vidal V, Srensen L (2001) Strategi og planlgning som lreprocesseks blde metoder, 2nd
edn. Handelshjskolens Forlag, Copenhagen
Chapter 6
When facing a complex planning problem requiring strategic decision making the
concerns initially relate to gaining insight into the nature of the problem and to
sweeping-in all aspects of relevance. This initial phase is of utmost importance as
it will influence all activities along the way towards making the best strategic
choice later on in the process by assessing the decision alternatives that have been
identified. Ideally we can consider the following main steps:
Scanning
Scoping
Assessing
Choosing
55
56
Core performance
Wider performance
Fairness
Diversity
Robustness
Hard methodology
Hard and soft methodology
Soft methodology
Hard and soft methodology
Hard and soft methodology
Below the five MOEs are reiterated with some consideration of their main use
when placed in a context of planning and strategic decision making:
MOE
MOE
MOE
MOE
MOE
1:
2:
3:
4:
5:
Based on Tables 5.1 and 5.2 each MOE is meant to inspire and guide exploration along one particular axis of insights; evidently the cognitive search in this
respect can in no way be finished. Therefore we have to speak of modes and not of
activities. As each mode in the tables is set out in a general way and not particularly for planning relating to complex strategic choices, we have to come to terms
with how the MOEs can become most cognitively effective when applied in a
context of planning and strategic decision making.
As concerns MOE 1 and MOE 2 goal seeking and viability together with
exploring purposes may become precise if seen as relating to core performance
and wider performance respectively. These formulations set focus on efficiency
and effectiveness as will be described below. MOE 3 and MOE 4 about fairness
and diversity respectively, can enter directly into methodology, while MOE 5 is
seen as having its focus on robustness, also to be treated further below.
With these modifications the five SP MOEs in Table 6.1 have been outlined,
which will guide the practical aspects of applying SP for complex strategic
choices.
The following sections will specify the methodology. First scoping will be
treated in Sect. 6.1 and afterwards assessing in Sect. 6.2. Both sections serve to lay
a basis for the treatment of the topic of choice intelligence addressed in Sect. 6.3.
Choice intelligence is coined in this presentation as a collective term for judging
more closely the overall effort and capability relating to decision support.
57
the process if such alteration is desirable for some reason. What matters, however,
in this early stage is that we do not exclude an option that could later onin case it
had been includedturn out to be a serious competitor among the alternatives to
come forward as the best choice. In a way then scoping is concerned with a kind of
boundary setting between what may loosely be described as included in the
decision space vs. excluded from the decision space. In a more mathematical
language we would describe this as omitting the fallacy of engaging in
suboptimisation.
Experience shows that sometimes scoping is paid too little attention as the
alternatives to be worked on more or less seem to present themselves. This may, of
course, be true in some cases. In complex, strategic decision making, however,
scoping becomes important as the boundary setting it represents is generally a
main element of the challenge represented by a complex plan problem.
On this basis all relevant modes of enquiry ought to come into play as they
explore our cognitive capability for what may facilitate the transformation of a
complex strategic decision problem into a set of choices that represent some initial
constraining of the decision space. Therefore this constraining should generally be
influenced by concerns of core and wider performance, fairness, diversity and
robustness, see Table 6.1.
Typically the strategic decision challenge arises on a background triggered by
either an opportunity or a threat or a combination of both. As it can be assumed
that the team addressing the strategic decision will have some preknowledge about
the possible core performance of alternativessome screening may be available
with rough estimatesthe wider performance and the fairness, diversity and
robustness issues come into the fore. Scoping therefore makes it reasonable to dig
deeper into the MOEs exploring such issues, and as already indicated in Table 6.1
this is the field of the soft methods in the toolbox.
Brainstorming (BS)
Mind mapping (MM)
Strengths, weaknesses, opportunities and threats (SWOT)
Critical systems heuristics (CSH)
Soft systems methodology (SSM)
Stakeholder analysis (STA)
Futures workshop (FW)
The first three methods are well known and practised in a number of more or
less formal versions (Leleur 2008). Brainstorming may range from free-and-open
discussion to a version based on rules, where a facilitator conducts the session that
will typically contain a sequence of questions; mind mapping is also relatively
58
well known, where ideas and especially how they interrelate are brought forward
successively as the process goes on and ends up presenting the team involved with
what is sometimes called rich pictures. Especially wider concerns and diversity
issues can be shed light on using brainstorming and mind mapping in combination.
A more structured way of thinking about the complex, strategic problem can be
obtained by a SWOT analysis, where internal and external factors are approached
by imagination and consideration of respective strengths versus weaknesses and
opportunities versus threats leading to a SWOT-matrix that can facilitate the
further process of scoping the alternatives.
What constitutes the differences between the methods is their balance between
being unstructured and thereby allowing open discussions and being more structured and thereby securing a relevant result (in some respect) to come out of the
efforts; needless to say that no right balance in this respect can be prescribed.
A concern always present in processes preparing decision making is the amount
of time and resources being consumed. If the team includesin addition to analystsa number of senior people from either the middle or maybe even the toplevel of the organisation this concern becomes even more outspoken. Typically
this will inhibit the use of methods more demanding in time than brainstorming,
mind mapping and SWOT. However, there may even on this background be good
reasons to apply a more demanding soft method that has shown a capability to dig
out knowledge about the decision problem in hand and have a critical influence on
the outcome of the process. In the systemic toolbox the available methods are
critical systems heuristics (CSH), soft systems methodology (SSM), stakeholder
analysis (STA) and futures workshop (FW). Of these four methods the first two
will be described below with regard to their potential. The latter two methods are
less demanding in prescribed content but not in time, and they are treated
explicitly as part of the demo-case in Chap. 7. As this is not the case with regard to
CSH and FW, the following two subsections will concern CSH and SSM, both of
which have a strong record of practical applicability, for which reason they are part
of the SP toolbox.
59
The intention behind the CSH tool is that all 24 questions consisting of the 12
is-questions and the 12 ought-questions should be carefully dealt with.
However, in practiceevidently depending on the actual problemit is the
experience of applying CSH as part of the SP framework that it is perceived as
cumbersome (at least by some people participating in the team) to pay the method
full respect in this way. Therefore an alternative way of making use of CSH is to
concentrate on the questions that seem to be most relevant and productive with
regard to obtaining new insights concerning the particular problem dealt with.
However, CSH used in this selective way can also be really worthwhile.
60
activities nos. 1, 2, 5, 6 and 7 are embedded in what Checkland terms the real
world, while nos. 3 and 4 are seen as systems thinking about the real world. The
structure described is shown in the upper part of Fig. 6.1.
The principle of SSM can be understood by surveying the stages of the
methodology. Stages 1 and 2 try to build as rich a picture of the problem
situation as possible. This implies collecting several perceptions. Specifically, it
has been found relevant to investigate both a slow-to-change structure and a
continuously changing process. By relating structure to process, essential
characteristics of the situation may be revealed. The function of the initial
stages is to obtain an expression which can serve as a background for relevant
choices.
The subsequent stage 3 concerns what Checkland calls Root Definitions. Their
purpose is to define one or more relevant systems in a way that makes it possible to
discuss their nature more openly. Such definitions constitute a survey of the
problem situation and provide the base from which such a survey and its implications can be further developed. It should be emphasised that we are not dealing
with real-world problems in stages 3 and 4, but with intellectual constructs or ideal
types. In this way, each Root Definition (RD) or set of Root Definitions is an
abstract ideal type of a purposive system that Checkland calls a Human Activity
System.
In stage 4, what Checkland calls Conceptual Models is formulated. This can be
done by simply bringing different verbs together, but more formal systems rules
formulated as part of the methodology can also be applied. Conceptual Models can
be seen as structured sets of activities combined logically in accordance with their
underlying Root Definitions. Thus, these models are not in the real world either,
but are a refinement or further building up of the ideal thinking.
The aim of stage 6 is to make use of the comparison results obtained in stage 5
to discuss possible, relevant changes. These should be both desirable on the basis
of the insights from Root Definitions and Conceptual models and they should also
be culturally feasible in the actual context.
In stage 7, action should be taken on the basis of the outcome of stage 6,
whereby the learning cycle is closed and a new situation obtained.
The outlined SSM stages can be based on further explanation. Among other
things a mnemonic CATWOE has been devised that can function as a kind
of checklist for further considerations. Each letter indicates a type of question
or consideration to be reflected upon, for example W for Weltanschauung,
where the German word for world view has been chosen to indicate a certain view behind a specific Root Definition and its associated Conceptual
Model. The other letters in the mnemonic are explained in the lower part of
Fig. 6.1.
Next, in stage 5, a comparison is carried out between the findings from stage 2
and the suggestions derived from the Conceptual Models in stage 4. Now we are
dealing with real-world problems once more by asking what features of the
Conceptual Models are especially different from present reality and why.
Problem
situation:
Unstructured
61
Taking action
Action
Changes:
Desirable and
feasible
Problem
situation:
Expressed
Comparison
Real world
Systems thinking
about the real world
Conceptual Models of
the system concepts
named in the Root
Definitions
Root Definitions of
relevant Human
Acticity Systems
Amplification
(1)
Ownership (O)
Ownership of the system, control, concern or sponsorship; a wider system which may discourse about
the system
(2)
Actors (A)
(3)
Transformation (T)
(4)
Customer (C)
(5)
To this list is added the sixth item which is, by nature, seldom if ever explicit in a root definition but is always implicit and always relevant:
(6)
Weltanschauung (W)
The (often unquestioned) outlook or taken-forgranted framework which makes this particular RD a
meaningful one
Fig. 6.1 Soft systems methodology (SSM): Process and a mnemonic checklist (CATWOE)
for particular considerations. Adapted from (Checkland 1985, p. 19)
62
We can sum up the meaning and potential of SSM the following way (Leleur
2000, p. 201):
It is important not to restrict the planning unintentionally, that is not to impose
any kind of unwanted closure
Basic choices with regard to concepts (Root Definitions) and their use in
modelling (Conceptual Models) fundamentally bias the whole planning process,
and
Radical considerations may be made possible if alternative ideal constructs are
cultivated and confronted with an expressed real-world problem situation.
The examination of core and wider performance will primarily draw on the
functional and the interpretive MOE respectively.
Generally the core performance will comprise economic issues such as gains
termed benefits and expenditures termed costs. The wider performance of a choice
alternative will concern other issues that cannot be treated in money terms.
63
Therefore in SP, core and wider performance are approached as economic and
non-economic assessments respectively.
Not surprisingly economic assessment in organisational decision making is a
prominent issue and rightly so as organisational development (or in a more pessimistic mode: organisational survival) depends on a combined outcome of decisions made (strategic, tactical and operational) that at the end of the dayor better
in an accounting perspective: end of the yearshould consist of black figures on
the bottomline instead of red ones. What complicates this matter when focusing on
strategic decision making is that often it is not possible to obtain an economic
assessment that covers the attractiveness of the decision choice alternatives in a
satisfactory way. With complex strategic choices we typically face the challenge
that matters of importance are non-economic.
In the SP framework this is treated by assessing the core performance by use of
cost-benefit analysis and the wider performance by use of multi-criteria analysis.
Both methods belong to the hard methods in the systemic toolbox. To see them as
analytic (requiring a functional MOE) and semianalytic (drawing also on the
interpretive MOE) respectively, will follow from the way they are described in the
subsections below. As concerns the other hard methods, scenario analysis and
preference analysis will be part of the demo-case description in Chap. 7, whereas a
description of risk analysis as applied in SP will follow after the subsections below
about cost-benefit analysis and multi-criteria analysis.
64
The second quotation has a focus on the short-comings of market pricing and
perceives MCA as an engineering approach in contrast to an economics approach.
In a comprehensive presentation of MCA methods for regional planning from
1988, it is stated that:
65
there exists the situation where the market price mechanism is not any longer well
functioning and for which alternative evaluation criteria have not yet been well established. The market price mechanism combined with the efficient allocation of resources
has not worked as the proper evaluation index for planning. This problem is known as
market failure. A major subject of MCDM (multicriteria decision methods) research is
thus to resolve the theoretical evaluation problem. this research highly intends to
take problem-solving as well as problem-findings aspects into major consideration: thus
this is an engineering approach in contrast to an economics approach. (Seo and
Sakawa 1988, p. xiii).
Hence, MCA methods are informed by preferences that are available from
decision makers. For this reason MCA accommodates an open-ended process
much better than is the case with CBA. One may say that in this way CBA takes on
a tinge of objectivity compared to MCA. The purpose of MCA is therefore not to
find some kind of correct, hidden answer but rather to assist the decision makers in
mastering the (often complex) information involved and advancing towards a
solution (Gissel 1999; DMG 2010).
Two main branches of MCA methods have been found particularly useful for
assisting decision making regarding complex strategic choices. One concerns
using multi-attribute utility theory and is represented in the toolbox with SMART
(simple multi-attribute ranking technique). This type of method consists of scaling
and weighing the different attributes of the alternatives to achieve the one which
scores the highest. The other branch proceeds by applying pairwise comparisons,
which has been found useful in the way decision makers can be involved in the
assessment. This methodology is represented in the toolbox by AHP (analytic
hierarchy process). AHP is by far the most wellknown of the pairwise MCA
methods. Another method is REMBRANDT (ratio estimation of non-dominated
alternatives); REMBRANDT can overcome certain difficulties that may arise in an
AHP session. The practical use of the SMART and AHP methods is presented in
Chap. 7, and further information about SMART, AHP and REMBRANDT is given
in Appendices A and B about the COSIMA and SIMDEC methodologies.
66
that the employees and management of the company TRANS-IT Consult have
different opinions, which may have an influence on what is the best relocation
decision.
Uncertainty may hamper the expected outcome of a decision. Therefore complex strategic choices clearly involve various types of risks. The higher complexity
involved due to the open-ended type of change earlier seen as characterising the
strategic decision (with tactical and operational decisions seen as related to closed
and contained change, respectively) means that the risks involved are important. In
the SP framework risk analysis (RA) by use of Monte Carlo simulation is included
as one of the hard methods in the systemic toolbox. But as will be underscored in
what followsmainly in Chap. 8this type of calculation has to be accepted with
caution. Uncertainty and risk analysis is also treated in Appendices A and B about
the COSIMA and SIMDEC methodologies linking, respectively, cost-benefit
analysis and multi-criteria analysis in COSIMA and multi-criteria analysis and risk
analysis in SIMDEC.
67
68
6.3.2 SP as Teamwork
In Chap. 4 the systemic process leading forward towards a situation where a
decision about complex strategic choices could be made was related to the
building of competence, and in Chap. 5 a systemic toolbox was formulated with
individual methods and techniques categorised as being either hard or soft. What
characterises the use of formal models and group deliberations? Gilboa has raised
the issue of how unqualified decision making can be avoided:
the use of formal models may be of great help working in groups and brainstorming
often helps. This should be qualified, because group decisions are not always better than
individual decisions. Groups that differ in their motivation may find it hard to make
coherent decisions, and if they do, the decisions may be very conservative, and may also
be swayed by charismatic personalities. But individuals who discuss a problem together
and then go their own ways to make individual decisions will generally make better
decisions than they would on their own. Groups tend to be better than individuals in sheer
analysis, with many ideas being brought up, challenged by others, compared, and analyzed. (Gilboa 2011, p. 20).
69
Even acknowledging that highly successful strategic decisions may have been
taken by one single person with a short time to prepare it, it is the idea and message of
this book that for important complex decisions it is worthwhile to spend time and
accumulate a certain amount of efforts for preparing decision making. At the same
time it is worthwhilewith the way the SP framework has been set upto organise
the strategic decision making as a group effort. Therefore SP is best pursued as
teamwork as will also be the case in Chap. 7, which presents an example of company
relocation. For a company to relocate its headquarters, this type of problem is
certainly an issue involving a situation with complex strategic choices.
Main points and findings of this chapter
When facing a complex planning problem requiring strategic decision making
the initial concerns relate to gaining insight into the nature of the problem and
to considering in principle all aspects of relevance. With this as our point of
departure two major activities are scoping and assessment. Scoping aims at
determining a preliminary set of choice alternatives, whereas assessment aims at
identifying the most attractive of the choice alternatives.
In the scoping mainly the soft methods in the systemic toolbox assist the
planners in their deliberations, while mainly the hard methods are applied in the
assessment for the determination of the consequences and risks that relate to
each of the choice alternatives.
Scoping and assessment are necessarily interrelated activities. What matters in
scoping is that an option or choice alternative is not excluded if it later on in the
assessment could have come forward as a serious competitor to the alternative
assessed as being the most attractive one. Therefore the scoping should be
returned to and reconsidered on the basis of the assessment.
All the 2 9 7 methods in the toolbox have their particular functions and relate in
different ways to the five MOEs behind their inclusion. As the later case
descriptions bring forward they are used in combinations that were found
suitable in the specific study.
Use of group processes arranged as decision conferences and making use of soft
and hard methods in combination are major characteristics of systemic planning.
Finally in this chapter some evidence is given that the blending of methods,
known as multi-methodology, and making decisions as teamwork are both
effective and beneficial for the end result. The validation of SP as a decision
support approach is carried out after the detailed demonstration of one possible
use of SP in the following Chap. 7.
References
Checkland P (1981) Systems thinking systems practice. Wiley, Chichester
Checkland P (1985) The approach to plural rationality through soft systems methodology.
Lecture Notes in Economics and Mathematical Systems 248, Springer.
70
DMG (2010) Decision modelling group compendium decision support: theory and practice,
DMG. DTU Transport, Technical University of Denmark, Copenhagen
ECMTEuropean Conference of Ministers of Transport, Group of Experts (1981). Exchange of
information criteria applied to transport infrastructure projects. ECMT Publication, Paris
Gilboa I (2011) Making better decisions: Decision theory in practice. Blackwell, Malden
Gissel S (1999) Decision aid methods in rail infrastructure planning, Ph.D. thesis, Report 4,
Department of Planning, Technical University of Denmark
Jackson M. C (2000) Systems approaches to management. Kluwer Academic, Berlin
Leleur S (2000) Road infrastructure planningA decision-oriented approach, 2nd edn. Polyteknisk Forlag Lyngby, Denmark
Leleur S (2008) At navigere mod fremtiden: systemisk planlgning som ide og metode.
Polyteknisk Forlag Lyngby, Denmark
Mingers J and Gill A (1997) Multimethodology. Wiley, Chichester
Mitchell S.D (2004) Why integrative pluralism?, emergence: complexity and organization,
6 (12):8191, ISCE Publishing
Neisser U and APA Task Force members (1995) Intelligence: Knowns and Unknowns, report
from the American Psychological Association (APA)
Seo F, Sakawa M (1988) Multiple criteria decision analysis in regional planning. D, Reidel
Publishing Company, Dordrecht
Ulrich W (1983) Critical Heuristics of social planning. Wiley, New York
Chapter 7
71
72
taken soon. There are a number of requirements to be met by the new site and
buildings ranging from a restructuring of offices into an office landscape on the
basis of a just completed, although minor reorganisation to including also the
concern that the new location should be in line with the companys image. Among
a number of other requirements are also that the economic consequences should be
sound, which means that although relocation will open the door to new opportunities, the economy is not without significance. Another requirement concerning
the relocation decision is that the staff emphasises that it must not be too difficult to
get to work. A number of factors are thus to be taken into account and some of
these are not readily measurable. A move from the present central location in
Copenhagen close to a major transport terminal should in general with all pros and
cons taken into account be considered advantageous.
The management decides to set up a group with representatives from both
management and the various departments and staff groups in order to prepare a
basis for the final decision. A newly employed engineer who has eagerly participated in the discussion and has raised the issue at the companys annual meeting
is invited to become a member of the group as an input from fresh eyes is
welcomed by the CEO of the company. Accepting this invitation he proposes that
a systemic planning (SP) approach may be useful, which he suggests on the basis
of his recent attendance in a class in planning theory at the Technical University of
Denmark. His proposal is prompted by the numerous factors and uncertainties he is
already aware of. It is agreed to give SP a chance as, by considering the long-term
consequences for the company, it seems reasonable that factors other than the
purely economic ones should influence the final decision about the relocation site.
73
On this basis the work of the Move Group falls into three phases: (I) they will
discuss the commissioned consultants report when available, (II) they will then
work to mark areas beyond the economy which should influence the decision, and
(III) they will conduct a final decision conference to establish their recommendation for the decision to be taken.
74
Denmark:
D1: Tietgens Hus Copenhagen
D2: Arne Jacobsens All restaden
D3: Torvevej Vallensbk
D4: lholmparken Hillerd
Sweden:
S1: Vstre Hamnen Malm
S2: Hyllie Centre Area Malm
S3: Svgertorp Malm
S4: Sder Helsingborg
Fig. 7.1 The eight alternatives for the relocation of TRANS-IT Consult
75
Alternative
B/C rate
Tietgens HusCopenhagen
Arne Jacobsens Allrestad
VallensbkCopenhagen
lholmparkenHillerd
Vstre HamnenMalm
HyllieMalm
SvgertorpMalm
SderHelsingborg
1.00
0.80
2.40
1.25
2.07
3.05
2.83
3.64
TRANS-IT Consult should be somewhere in the resund Region, but that within
this constraint no site which ought to be considered must be left out.
Next the Move Group address the economy calculations of a cost-benefit type,
which have been created by the consultant. Based on a wealth of information about
rent levels, operating costs, removal costs, efficiency gains through redeployment
and costs associated with breaking up (disruption costs) and differences of Danish
and Swedish wage levels, etc. B/C rates have been established, which express the
attractiveness of each alternative from an economic viewpoint. These results
represent the core performance according to SP terminology and are shown in
Table 7.1.
Specifically, this table indicates that in economic terms a location in Helsingborg
will be preferable. Since almost all rates are above 1, the B/C calculations also show
that the very decision to move seems economically correct.
The B/C-rates give rise to intense discussion of, among others things, what is
currently included and what is yet to be taken into account. As concerns the latter
it is furthermore discussed whether these matters are possible to include by
calculation or only by being addressed by wording.
After the work programme earlier agreed upon, the next phase concerns getting
more clarity about the influences SP addresses as non-economic and which relate
to the wider performance of the alternatives. In reality it is hard to separate
economic and non-economic factors. One of the factors that has already, not least
from the staff side, been given much attention is the accessibility of the new
location compared to the existing location, which is quite close to commuter trains
and subway at Nrreport Station in central Copenhagen.
This situation could well be a restarting of the cost-benefit calculation as there
are well-defined transport costs and an established practice of valuing the savings
and costs related to changed travel time. There are also ways of calculating the
effects of more or fewer shifts and for waiting-time in this respect. Depending on
the specific alternative some home-to-work travelling may change to car travelling
(the present location is very accessible by public transport). This effect could be
calculated in economic terms comparing the difference between the before and
after situation as it is called in cost-benefit analysis. Actually using a geographic
76
information system (GIS) and applying the knowledge of staff residences fairly
accurate calculations become possible.
The budget-responsible person of the Move Group, however, realises that a new
extended cost-benefit analysis conducted by the management consulting firm with
inclusion of changed daily travel patterns of the employees will cost so much that
there will be no budget means left for the planned final decision conference.
However, based on a suggestion from the newly hired engineer that a less costly
approach can be used when considering also the change in travel pattern, it is
decided to stick to a work programme with a final decision conference.
Actualised by the home-work accessibility issue the Move Group asks whether
there are other methods in the systemic toolbox that could be used to gain insight
into the non-economic factors considered to influence the choice of location.
It appears that in addition to brainstorming more demanding methods exist, which
can be used to pursue the issue. The methods in this respect are soft systems
methodology (SSM) and critical systems heuristics (CSH). These methods are very
suitable if you start from scratch. But the Move Group finds that it is actually not
the situation anymore. There have been until now three sessions, and there is a
feeling in the Move Group that both phase 1 and phase 2 of the work programme
have been covered reasonably well so it is time to prepare for phase 3 with the
decision conference.
77
78
In this respect it is important that the conference facilitator initially explains the
underlying decision model theory to the participants. This needs not be with all
details set out but in a way so the participants obtain an overall understanding of
how the process is organised and why the different steps taken are important. In
brief there should be no black boxes as they will detract from the purpose of a
common learning that should lead to a good decision. Therefore the decision
conference should appear as a natural way of unfolding the complex decision
problem without having to bother too much about technical details. The description below of the decision conference conducted to obtain a best choice about the
new company location in the resund Region seeks to demonstrate that this can be
achieved.
79
This instigates a larger discussion, and it becomes relevant to draw on the views
that were formulated already in phases 1 and 2. These can now be reconsidered
and further elaborated on to form a set of criteria. The facilitator points out that the
criteria should not overlap and that they will function to represent in principle all
factors of relevance for the relocation decision.
After some deliberation the participants decide that the following criteria, or
effects, should be included as they are defined below.
Proximity to customers: The companys position in relation to its primary
customers
Image: The image the location and the building(s) present, especially to
customers
Office size: The size of the new office environment (m2/person) and the layout
flexibility
Site facilities: The facilities associated with the site and the surrounding area
Parking facilities: Parking facilities for cars, etc.
Public transport: Its accessibility by public transport (bus, subway, train)
Individual accessibility: Its accessibility in relation to individual transport
(car, etc.)
Global accessibility: Its accessibility in a global perspective in terms of proximity to airport
At this stage it is important to emphasise that the list only highlights that the
mentioned criteria are expected to influence the choice of the relocation alternative, not how important each criterion is in comparison with the other criteria on
the list. Therefore the order of presentation of the criteria is not relevant as it does
not reflect a relative importance associated with the criteria. Including also the
CBA assessment from the management consultants report an overview of decision criteria can be set out in a decision tree introducing two intermediary levels
between the relocation choice problem and the criteria that have been formulated,
see Fig. 7.3.
Taking a look at the decision, it is essential to consider whether relevant criteria
are missing and that overlapping between the criteria does not occur and that they
can be regarded as truly additive to the effects in the cost-benefit calculation. The
facilitator here tells the participants that a preliminary decision foundation has
been established consisting of what is commonly referred to as the cost-benefit
analysis (CBA) effects and the multi-criteria analysis (MCA) effects. From a
methodological point of view she states that some major questions remain, namely
how important the MCA effects are seen relatively to each other and how
important the group of MCA effects combined is compared to the group of effects
included in the CBA. She introduces the term trade-off between CBA and MCA in
this respect as one of the upcoming tasks. First, however, it will be necessary to
address the MCA effects more closely. This concerns what is called the scoring of
each relocation alternative under each of the MCA criteria/effects. Characteristic
here is that for each MCA criterion a value function (VF) is defined so that the best
80
Fig. 7.3 The formulated decision tree for the TRANS-IT Consult relocation case
alternative is assigned the value 100 and the lowest the value 0. The other alternatives are graded linearly so they obtain scores between 100 and 0.
The technique made use of here is the analytic hierarchy process (AHP) (Saaty
1977). Under each of the eight MCA criteria all the eight alternatives are compared
two at a time. This process is supported by the smartboard screen and facilitated
with regard to software prompts by the analyst, and for each MCA effect in the
decision tree the participants are asked how alternative A performs compared to
alternative B under this criterion. The A versus B performance (A and Bs
attractiveness in accordance with, for example, proximity to customers) can be
better, equal or worse (the latter corresponds to B being better than A).
81
The essence of this scoring is that all preference information collected from the
group in this way can mathematically produce the VF scores for all alternatives.
For each of the eight MCA effects eight scores are thus produced, with one of these
being 100 and one 0 and the remaining scores in the range in-between. Evidently
what matters here is how the preference information is graded by the group, and
for this purpose a semantic scale is used. The grading here is explained by just
paying attention to A being better than B (as A being worse than B can be
transformed into B being better than A).
With A being better than B the group has to decide on the following semantic
gradings (Saaty 2001, p. 73):
Equal importance
Moderate importance
Strong importance
Very strong importance
Extreme importance
These are the five main gradings that the group initially has to consider. In case
that doubts are expressed about which of two neighbouring grades that best
expresses the intensity of importance, in-between gradings can also be used as
indicated below with grading-numbers also shown (Ibid.) (Note that the gradingnumbers are treated in Appendix A. Here it suffices to observe the semantic
presentation of the various intensities that can be made use of by the group
members in a pairwise comparison).
Equal importance
In-between grading
Moderate importance
In-between grading
Strong importance
In-between grading
Very strong importance
In-between grading
Extreme importance
(1)
(2)
(3)
(4)
(5)
(6)
(7)
(8)
(9)
The facilitator informs the group that their promptsassisted by the analyst
feed into a numerical method that calculates the VF scores. Furthermore, asked by
one of the participants about interdependence, she explains briefly about the
concept of transitivity simply meaning that if football team I beats team II and
team II beats team III, then we should expect team I also to be able to beat
team III. What makes football entertaining is that it is not always so! In scoring,
however, she explains you have to observe consistency in the grading to some
extent so transitivity is not violated. However, she adds that the grading in fact by
answering all possible pairwise comparisons produces a kind of surplus information that helps stabilising the outcome so that minor inconsistencies can be
handled. Furthermore, she notes that the software has a capability to indicate
82
83
Fig. 7.4 The scores for the eight alternatives assessed by eight MCA effects
Some issues are raised but turn out to be explainable. In fact some notes taken
by the subgroup help shed light on the image scores and the public accessibility
scores, and the underlying preference gradings that can be recalled on the screen
are accepted by the whole group. It is decided that in the reporting of the work
of the Move Group a kind of log book will be added as an appendix giving
others the possibility to look into gradings that they consider to be of main
interest for the decision making. Fortunately both the company management and
the employees are represented in the Move Group but there is a general
agreement that log book information is important and would have been even
more so, had not both main stakeholder groups been represented in the work.
The facilitator adds that had the work been conducted mainly as consultant work
supplemented with maybe a few persons from the company as informants, the
necessity of producing information about the critical parts of the process would
evidently have grown in importance.
In this way each alternative is scrutinised, and it is found that the scores seem to
represent the alternatives pretty well. The facilitator notes that the columns for
each alternative cannot just be added to point out the most attractive site as this
would reflect equal importance of each effect; however, she adds that the scores to
some extent indicate which alternatives can end up being the most interesting to
concentrate on when the final decision is to be taken.
That the MCA effects are not equal in importance is the background for the
fourth question asked by the facilitator:
Question 4. Can you agree about an order of priority for the MCA effects as
concerns an assessment of their relative importance?
84
ROD-weight
1
2
3
4
5
6
7
8
0.23
0.20
0.17
0.14
0.11
0.08
0.05
0.02
Image
Proximity to customers
Global accessibility
Site facilities
Individual accessibility
Public transport
Parking facilities
Office size
Individual accessibility
Public transport
Site facilities
Parking facilities
Office size
Image
Global accessibility
Proximity to customers
The facilitator briefly explains why this particular question is asked. In fact she
could have asked the participants directly as a group to enter criteria weights summing to 1. However, experience shows that such a process when viewed group
dynamically is quite complicated. As was the case with the previous scoring of the
alternatives, pairwise comparison could also be a possibility to set the criteria
weights. However, she states that another approach will be used. This concerns a
ranking approach which can better accommodate iterative group deliberations due to
the more subjective nature of weight-setting as compared to the more objective
comparisons of the alternatives made to assess their scores. The technique she has in
mind is rank order distribution (ROD) weights that by mathematical principles and
probability theory can provide a good estimate of the weights that lie behind a given
ranking of the MCA effects (Roberts and Goodwin 2002). The facilitator then sets
focus on whether a joint prioritisation of the MCA effects can be achieved.
However, this happens not to be the case in this group of decision conference
participants. The ranking of non-monetary MCA effects thus depends on whether
the company management representatives set the rank order of criteria or whether
the employee representatives do. Therefore it is decided to continue with two
rankings, namely a ranking or priority order devised by the company management
and a ranking devised by company staff.
Having recognised that this was the best way to move forward, the Move Group
splits into two subgroups to focus on the respective rankings without undue
interference. However, before doing this the management representatives briefly
note that at the end of the day the decision belongs to the company, which makes
the employees respond that when raising the issue for the first time some months
back the managing director had emphasised the necessity to take all interests into
account. In this way Table 7.2 is produced showing the two different rankings and
their associated ROD weights.
The two rankings can be viewed as an expression of two different attitudes or
sets of preferences towards the relocation decision. Thus the table shows that
management will prioritise criteria such as image, proximity to customers and
global accessibility. These are primarily concerned with the companys public face
and turnover, which are obviously very important to address for the management.
In this context, it then becomes less important to consider public transport, parking
and the size of the offices.
85
The staff, however, wishes to put great emphasis on individual and public
transport accessibility and the facilities at the location. This is not surprising as
these criteria have a major influence on their daily working conditions. Less
important for the staff is global accessibility and proximity to customers.
86
which to some participants is smaller than they earlier conceived it to bethe two
stakeholder groups, however, can agree that Arne Jacobsens All is the most
attractive alternative and Sder, Helsingborg, the least attractive.
Were the final decision to be based solely on the non-economic factors and thus
ignoring the economic factors presented in the management consultants report
with the CBA results, the Move Group would be pretty close to having a foundation for a location decision. An interesting fact is, however, that the CBA
component suggests a solution that is the opposite of the one based on the MCA
component of the overall analysis as Arne Jacobsens All has the lowest B/C rate
of all the alternatives, see Table 7.1. Furthermore the least attractive location
choice according to MCA Sder-Helsingborg is when based solely on the CBA the
most attractive. This accentuates the need to consider CBA versus MCA, which is
the concern of the CBA versus MCA trade-off analysis.
The facilitator presents the principles to be applied in the CBA versus MCA
trade-off analysis, and the analyst contributes with some relevant screen shots to
illustrate how to make use of both the CBA and MCA component parts at the same
time. More technically this concerns how to weigh the two parts together.
A basic principle here is that the results of the CBA part should not be changed,
which means that the calculated B/C rates must be maintained. What is being
changed is the MCA influence on the overall attractiveness relative to the CBA
influence, see Fig. 7.6, which serves to describe the meaning of the trade-off
parameter named MCA%.
As CBA is kept fixed, what varies is the MCA%. This parameter as explained by
the facilitator can be understood in an intuitive way so that a high MCA% means a
high influence on the overall attractiveness assigned to each alternative dependent
87
Fig. 7.6 Illustration of the MCA% in the CBA versus MCA trade-off analysis
on its MCA-scores profile, whereas a low MCA% means less importance paid to the
MCA component. If an even influence is sought the MCA% is set to 50%; a very
high influence can be 80% and a very low one can be, for example, 10%.
On this basis the facilitator asks the following question:
Question 5. How do you perceive that CBA and MCA should affect the
overall result expressed by an appropriate MCA%?
As the ranking of MCA effects could not be reached based on consensus about a
single priority-order listing of criteria, it is conceivable that a single pre-defined
MCA% cannot be agreed upon either. Relatively soon the participants request to
see the consequences of a variety of chosen MCA% values, which then could be
the point of departure for further discussion. Agreeing to proceed in this way, the
facilitator, however, asks each participant to write down a value (not at this stage
to be shown to the other participants), which can be used as background for the
upcoming further discussion.
Furthermore, it is agreed to continue both with the management and the staff
MCA results. Here a large screen or even a number of screens are advantageous as
they allow results from the two decision strategies to be studied simultaneously.
First to be shown on the screen are the results according to the management
strategy. The result is indicated as a value composed as a contribution from CBA
and from MCA and is described as the total rate of return (TRR) value. Thus TRR
for a given MCA% expresses the overall attractiveness based on a CBA and an
MCA contribution. The background for expressing TRR as a rate is that MCA
88
Fig. 7.7 Management decision strategy: total rate of return (TRR) based on applying both CBA
and MCA for different values of the trade-off parameter MCA%
(similar to CBA) is considered also by taking costs into account. Therefore TRR
comprises the value of the economic (monetary) and non-economic (non-monetary) effects when compared to the total cost invested in the relocation of the
company from its location in central Copenhagen to the studied location alternatives. From the first screen that appears with all eight alternatives it is possible for
the participants to narrow down and select, for example, the four most interesting
alternatives, which has been done in Fig. 7.7. For the MCA% equal to 0 the CBA
result is shown on the TRR-axis; as MCA is weighed in with still higher MCA%values, the relative performance of the alternatives change as indicated by the
curves.
It can be seen that the TRR of the individual alternatives depends on the actual
value of the MCA%. In the interval from around 0% and until around 25% Sder
in Helsingborg is the most attractive alternative, which is in accordance with the
result of CBA. From around 25% and until around 45% the picture changes so
Hyllie in Malm becomes the most attractive alternative. Note also that in this
interval Arne Jacobsens All changes from being the least attractive alternative
and from around 45% it becomes the most attractive one for relocation.
This development is in agreement with the result of the MCA, which is more or
less the opposite of the CBA result. This is not surprising as alternatives seen as
attractive in the MCA will gain in overall performance when a higher MCA%,
which means a stronger leaning towards the MCA assessment, is applied. The
change in MCA% is also to be seen as a decrease in the consideration of the core
performance associated with CBA towards an increased influence from importance
of the alternatives wider performance associated with the MCA. Note that an
alternative as Helsingborg with low scores on the MCA effects is only moderately
adjusted with a higher MCA%.
89
Fig. 7.8 Staff decision strategy: total rate of return (TRR) based on applying both CBA and
MCA for different values of the trade-off parameter MCA%
In a similar way the TRR values are calculated in accordance with the staff
decision strategy, see Fig. 7.8, also showing only the selected four most interesting
alternatives.
Figure 7.8 shows more or less the same tendencies as Fig. 7.7. Again the
participants realise that no profound differences occur between the two decision
strategies. What really happens is that the intervals of attractivity are slightly
changed. Helsingborg is still the most attractive alternative between 0 and 25%,
and Hyllie now becomes the most attractive between 25 and 55%; then again Arne
Jacobsens All is the most attractive alternative for the higher MCA% values.
The difference in outcome as regards the management and the staff decision
strategies are due to the different priorities for the relative importance of the MCA
effects. The facilitator notes that greater differences could occur given that other
distributions of MCA scores had been found. Nobody, however, wants to reconsider the scoring made earlier.
The MCA% to be decided as recommendation for the final decision making
ought to depend on the type of decision problem. It will be relevant here to give
examples to the participants. For a large infrastructure investment maybe a
maximum of 30% would be relevant, but in the actual case higher values could be
applied as the relocation decision for the company may be the most important for a
long time. As a rule it can be stated that the uniqueness of the strategic decision
matters. This information starts a discussion among the participants about not
choosing a too small value. At the same time issues turn up about general economic development (regional, national and international) relating to the companys markets and otherwise. This leads to a wish to engage in further
consideration about taking possible alternative developments into account and also
to engage in whether the decision to be taken will be robust.
90
91
techniques in the systemic toolbox, the futures workshop more explicitly helps
establishing a particular focus by running through altogether three phases: A first
phase with emphasis on being critical about the present situation, a second phase
with emphasis on outlining overarching visions for the future, and a third phase with
emphasis on the implementation of different alternatives of action.
In the decision conference the participants choose to set focus on the development
in the resund Region. Material is available from a large regional study (RIB
2007) that has, among other things, considered three developmental scenarios for the
regions urbanisation growth pattern towards 2045: (I) business-as-usual, (II)
monocentric growth, and (III) polycentric growth. In this respect a MCA% equal to
40 is chosen to serve as a point of departure meaning a close run between all the four
alternatives shown in Figs. 7.7 and 7.8 but with Hyllie as slightly more attractive
than the other.
It is recognised that in a region that is still only partially integrated, integration
between the Danish and Swedish sides of resund will continue to evolvein fact
one of the participants notes that quite a number of people from Malm on the
Swedish side of resund have actually been employed by TRANS-IT Consult in
recent years. Continued integration is expected to strengthen a monocentric
development around Copenhagen-Malm-Lund, which favours the performance of
Arne Jakobsens All compared to Hyllie. On the other hand a polycentric development may point to Hyllie as the best choice. At the same time the monocentric
development is related to a better overall economic development towards 2045
than is the case with the polycentric development. At this stage the Move Group
recognises that better economic development may point to increasing the MCA%.
The explanation is simply that the wider performance aspects associated with the
MCA effects become more important (we can better afford our vision about
TRANS-IT Consult and the image we want to build).
With a focus now on economic development and the impact this has on the
demand for consultancy and services offered by TRANS-IT Consult a background
is created for the last question to be asked by the facilitator:
Question 7. Do you find it relevant to make a risk analysis of selected factors
and parameters that have been made use of in the previous steps?
This question makes the group discuss the CBA results once again. What will
happen if the anticipated efficiency gains are only included as determined by
using various probabilities related to the different types of efficiency gains foreseen
to materialise with the relocation, and will this examination point towards this
being a critical issue? In the management consultants report these efficiency gains
were discussed but entered the CBA calculation as a set of rounded estimates.
What happens if the Danish-Swedish wage gaps fade away with time, also an issue
to be treated in the risk analysis? Will risk analysis change the view on the
individual location alternatives if, for example, the efficiency gains as estimated
92
93
competitor as the second priority proposal. The proposed alternatives are supplemented with comments and arguments that have been made in the light of the
progression of the decision conference.
Main points and findings of this chapter
The office relocation example demonstrates but one example of applying SP for
making complex strategic choices. Other methods could have been applied from
the toolbox, and the process could have been designed in another way.
This indicates that SP is basically an open-ended approach. Furthermore,
testing it on a diverse range of complex planning problems over the last five
years as described in Chap. 8 indicates a high degree of adaptability to the
specific study, which should be seen as a desirable feature.
In the theory chapters of the book (Chaps. 2, 3, 4 and part of 5) some efforts
have been made to pinpoint findings that can lie behind a reasonable practice for
supporting planning and strategic decision making. The SP theory skeleton
established in this way is, even in its present rudimentary version, seen as a
worthwhile and necessary effort as it functions as a backdrop against which new
ideas about process and methodology can be illuminated by the theoretical
works that were called upon over the chapters to specify the SP framework.
References
Barfod MB, Jensen AV, Leleur S (2008) STMA beslutnings-analyse: metoder, proces og
software. In: Carlsson CM (ed) Hllbart transport system fr inre och yttre attraktionskraft,
EU interreg project about strategic transport management in the resund region (STM)
MAH Malm, final STM report by Decision Modelling Group, DTU Transport. Technical
University of Denmark (DTU), Denmark
Goodwin P, Wright G (2010) Decision analysis for management judgement, 4th edn. Wiley,
New York
RIB (2007) Project resundsregionens Infrastruktur og Byudvikling, final report resundsregionen r 2045: Scenarier fr trafik och byudvikling. resundskomiteens Sekretariat,
Copenhagen
Phillips LD (2007) Decision conferencing. In: Edwards W, Miles RF Jr, von Winterfeldt D (eds)
Advances in decision analysis: from foundations to applications. Cambridge University Press,
Cambridge, pp 375399
Roberts RJ, Goodwin P (2002) Weight approximations in multi-attribute decision models.
J Multi-Crit Decis Anal 11:291303
Saaty TL (1977) Scenarios and priorities in transport planning: application to the Sudan. Transp
Res 11(5):343350
Saaty TL (2001) Decision making for leaders: the analytic hierarchy process for decisions in a
complex world. RWS Publications, PA
Chapter 8
95
96
97
8.1.2.1 Bypasses for Hng and Allerd, Denmark (2000 and 2006)
Assessing and prioritising Danish bypasses have initiated the work on methodology for making comprehensive assessments based on both economic and noneconomic impacts, which in the present context has been referred to as core
performance and wider performance. For a detailed description of this planning
problem the reader is referred to Leleur (2000).
Generally described the purpose of a bypass around a town is to improve the
level of service for the through traffic and to relieve the town of traffic. Core
performance is assessed by using a cost-benefit analysis (CBA) comprising
impacts such as improved (reduced) driving time and driving costs (petrol savings
among other things), while relieving the town of traffic is assessed by estimating
98
impacts relating to the improvement of noise and local emissions, etc., which is
handled by a multi-criteria analysis (MCA). The latter can also include improved
conditions for pedestrians and enhanced urban quality, etc.
No doubt the design of different bypass alignments, sometimes leading to
alternatives with quite different amounts of construction costs, and the selecting of
the best alternative make up a strategic study as the outcome can have long-term
consequences both locally with regard to, for example, urban development and
often also regionally and nationally by impacting on the trunk road network, etc.
The strategic decision regarding the bypasses depends on how much CBA and
MCA should count individually as the design solutions with their individual
alignment and design standard are engineering solutions that typically seek some
kind of compromise between the objectives inherent in CBA and in MCA
respectively. The methodology from the systemic toolbox is the composite model
for assessment (COSIMA), which builds on a linking of CBA and MCA.
In Appendix A COSIMA is presented in detail together with AHP, which is
used in the MCA part of COSIMA. It is demonstrated how COSIMA can be used
to organise strategic decision making in an interactive way drawing on decisionmaker preferences. At the same time it shows how different stakeholder viewpoints
can be treated with a preference analysis.
99
General concerns:
resund region one of several
spheres
The meaning of national
barriers
Drivers: market, clusters,
culture, etc
Infrastructure and development
Regional scenarios:
Economy, regulation,
transport, etc
Local integration versus
non-integration
Baltic Sea development: trade, etc
Competitive transport development
Specific concerns:
Limitations of cause-effect
model
Interpreting expressed
expectations
EU-wide scenarios:
Economy, regulation, transport
Trends: resources and technology
Trends: modal policies, etc
Ex-ante:
Local pro-coalition
Local environmental anticoalition
National interest
International pro-coalition
Ex-post:
National interest
resund region citizens
resund companies
International interest
100
limited an approach to deal with the complex consequences of the resund Fixed
Link (Leleur et al. 2000).
The systematic assessment pointed to an examination including both core and
wider feasibility based on CBA and multi-criteria analysis (MCA) respectively.
The results produced by COSIMA, see bypass case above, in a first-runbased on
combining some study assumptions, for example that high integration in the
resund Region would be obtained over the next decade, and that deregulation and
high economic growth would apply generally to Europeare shown below
(Leleur et al. 2004a, b, c):
CBA: Time savings, cost savings, accidents, etc
MCA: Network and mobility
MCA: Global emissions (CO2)
MCA: Employment
MCA: Logistics and goods effects
CBA and MCA:
0.85
0.11
0.02
0.32
0.05
1.35
The sole function of such first-run results is to give feedback to and guide
the further process. They provide, among other things, the interesting result that
the figures indicating benefit-cost rates (BCRs) as used in COSIMA, show no core
CBA-feasibility (0.85 \ 1) but a wider MCA-feasibility (1.35 [ 1, with CBA as
part of the result). In other words, the feasibility is dependent on factors that are
complex and uncertain.
Such a first-round examination lays the foundation for further unfolding of the
SP process. In the actual case where the implementation had taken place, further
analysis can assist in making decision support for possible follow-up initiatives
(regulation, other investments or interventions), but in other complex cases before
the primary decision about whether to implement the project or not has been taken
(an ex-ante study), a result where narrow and wider analyses point in opposite
directions will make it relevant for the planners to resume the process.
The case demonstrates the possibility of SP to produce scanning and assessment
information that can provide an overview of the many issues concerning the
decision about the fixed link. In transport planning the planning and deciding about
large infrastructure investments are among the highly complex problems.
101
No doubt the 3,000 m airfield is the ideal solution due to various considerations
including air technology and servicing (possible types of airplane, landing and
takeoff etc.), business, tourism, land-use with regard to developmental issues and
local environment (noise).
A decision tool was established to deal with the many economic and noneconomic issues. The tool showed that the 3,000 m alternative (from CBA
including also effects on Air Greenland) was basically not economically feasible
due to the high construction costs (low core performance), whereas both the 1,800
and 2,200 m alternatives could be defended economically but both did less well
with regard to the wider performance represented by the MCA effects included
(DTU 2007). In addition to the main study a master thesis (Mortensen and
Andersen 2007) found by combining COSIMA and CSH that the 2,200 m could be
a reasonable compromise. Specifically they conducted a stakeholder analysis that
confirmed that the 2,200 m solution would be a robust solution.
Similar to the previous case this case demonstrates that combining a number of
methods from the systemic toolbox can provide strategic decision making information that can explicate the design objectives behind the different alternatives in
a way that makes these comparable and accessible for structured debates across a
range of very different influences on the decision to be taken.
102
Many institutions are involved. Where should the final responsibility be placed?
With both civilian and military organisations involved will this give raise to
invisible barriers?
What does it really mean that the cross-border collaboration will be permanent
as compared to continued operation of two national units?
How is the regulatory EU legislation affected?
How are running, operational costs affected (economies of scale)?
For some staff members a colocation in one country will mean increased time
spent on commuting. How should this be considered?
Consideration of specific location possibilities
Identification of advantages related to cross-border collaboration
Constraints relating to issues of politics
The answering of these questions gave rise to a series of rich pictures (the
facilitator happened to be highly skilled at drawing and relating statements), which
produced the root definitions in SSM (where plain verbs express functionality in a
complex undertaking). This helped to form a practical range of options given
different attributes defined on the basis of the discussion among the participants.
The case demonstrates how soft systems methodology can assist problem
structuring. This can then lead to a range of options, which express the demands
from the stakeholders to the decision about the setting up of a common VTS. In
this case a number of methods from the systemic toolbox were made use of, but the
main experience here was obtained with SSM as a suitable tool to deal with the
complex problem in the first of two sessions.
103
104
Altogether 36 criteria are contained in this long list behind the strategic decision
making of this case related to choosing the best new light rail investments. As can
be noted some overlaps occur, which had to be dealt with. Furthermore, the
number of criteria had to be reduced to a practical number that should be less than
ten and at the same time with a minimum loss of important information.
This was taken care of in the decision conference as a number of impacts were
included in a CBA and a number of criteria included in an MCA.
The case demonstrates that it is possible to handle even long lists of criteria and
considerations. An important finding in this decision conference worth noting is
that the ranking according to the CBA turned out later on not to be affected by the
ranking produced by the composite use of CBA and MCA. To some of the participants this was really a surprise, but the facilitators explanation that MCA
incorporation need not alter the ranking of alternatives ended up by the participants to be perceived as a good result providing the interesting information that
core and wider performance may be correlated.
105
106
2010 round. The work on the ITS projects gave a possibility of going into depth
with the individual ITS projects.
As for the biking initiatives and projects also the work on the appraisal of the
ITS proposals had to be done without the traffic-economic knowledge that is
applied with ordinary highway projects such as the bypasses treated above. Even if
efforts to develop such knowledge are given high research and development
priority, it is doubtful whether a sufficient CBA foundation can be established for
these new types of projects. In the terminology of SP we can say that the noneconomic wider performance will continue to play an important role in upcoming
work. Not just in transport planning but also in other types of planning such as, for
example, energy planning with projects competing for funding but with lack of
sufficient data, this points to the need to develop assessment based on multi-criteria
analysis.
In Scandinavia, compared to mainland European countries such as the
Netherlands and France, there has been some reluctance to give up cost-benefit
analysis (CBA) and base transport decision making on multi-criteria analysis
(MCA). Some method proposals have suggested applying CBA as one of the
criteria in an overarching MCA. One prominent example here is the EUNET
methodology developed in a European strategic transport research with a number
of European countries represented in the research team (Tsamboulas et al. 1998).
However, the EUNET approach has not gained acceptance (authors personal
interpretation) because the CBA is transformed away from money into an index
value representing the CBA score.
On this basis the COSIMA (composite model for assessment) has been
developed, which has as its unique feature that the CBA result is maintained all the
time. Some responses to COSIMA have been that the approach is appealing as the
decision-makers can still keep their eyes on the CBA. Even when disagreeing
with the view that CBA information is more valuable for decision support than
MCA information, it is a point taken. What can be learned in this respect is to
objectivise the MCA procedures. In connection with both the biking and ITS
work, a log book documenting each pairwise grading with the arguments behind
has been kept. This means that each and every pairwise comparison can be
reconsidered and changed if new arguments are in favour of this.
As already noted COSIMA is treated in detail in Appendix A based on appraisal
of the Danish bypass alternatives for Hng and Allerd, which are towns on
Zealand located about one hour west and north of Copenhagen respectively. In
addition to COSIMA, based on combining CBA and MCA (Andersen and Petersen
2006; Leleur et al. 2007) another more recent method combining MCA and risk
analysis is treated in detail in Appendix B. The method called SIMDEC is
introduced below and is based on the assumption that decision-makers can handle
CBA information not just as a point estimate as in COSIMA but also as an interval
estimate that results from a risk analysis based on Monte Carlo simulation.
107
8.1.2.9 Fixed Link Between Elsinore and Helsingborg (2010 and 2011)
The case concerning the alternatives for the new fixed link between Elsinore and
Helsingborg is the most recent SP case. The methodology applied is based on the
SIMDEC approach, where robustness of socio-economic feasibility is one of the
decision criteria (Larsen and Skougaard 2010; Leleur et al. 2010).
Altogether four alternatives have been proposed and sketch designs have been
carried out to enter the further planning and political process. In addition to
specific purposes (for car or train only or for both) and to have preliminary
investment cost estimates, the sketched alignments, etc. make it possible to begin
an examination of the attractiveness of each of the alternatives. With a decision in
either 2012 or 2013, the new fixed link can be expected to open around 2020,
which will be a much needed relief of traffic from the current fixed link between
Copenhagen and Malm that opened in 2000.
This case demonstrates the use of risk analysis (RA) belonging to the hard
methods in the systemic toolbox. As appears from the description in Appendix B,
a new concept called certainty graphs can be used to provide the users with CBA
information including information about uncertainty relating to the estimate of
construction costs and to the prognoses made for the future traffic; the latter is also
referred to as the demand estimate. It can be noted that only these two elements are
included in the risk analysis. As mentioned in the appendix they have, however,
been found to be the main causes for large transport investment projects failing to
achieve a feasible socio-economic result in many cases (Priemus et al. 2008). The
type of risk analysis conducted is referred to as feasibility risk analysis (FRA)
(Salling 2008; Salling and Leleur 2009).
The perspective of developing SIMDEC is to have a methodology that can be
based on MCA and RA. A main issue here concerns how certainty graphs representing FRA-information will perform as decision-maker input in a decision
conference. Experience so far is promising based on, among other things, a
decision conference conducted in October 2011.
108
In their book Bhushan and Ray give a convincing description of how the analytic
hierarchy process (AHP) can assist the decision-makers when engaged in evolving
strategic decisions. They take their point of departure in describing what they see as
the generic decision making problem in the following way (Ibid., p. 11):
109
110
II.
As complexity is not a surface noise of the real that can be swept away, it
must instead be paid attention to by considering the fact that strategic realworld problems are expressions of open-ended change. In practice this
implies that learning can never be replaced by doing overall modellingbased optimising.
III. To deal with open-ended change and avoid unintended closure as regards
problem understanding it is necessary to apply different paradigms that less
formally but in a more practice-oriented way express different modes of
enquiry (MOEs). Altogether five different MOEs have been identified to
produce what was termed a cognitive billboard having the purpose of
bringing cognitive awareness and alertness to SP typically early on in the
process.
IV. The systemic process can be organised in different ways, however, always
with the purpose of triggering search-learn-debate activities around the
complex plan problem. In principle one person could undertake this (with
debate then becoming reflection) as a desk study but there are obvious reasons
for setting a team covering a wide range of competencies and skills. Furthermore there are good reasons to engage a moderator, coach or facilitator.
The latter is the term used when organising the process as a decision conference (see below).
V. The systemic toolbox has been developed to contain a number of hard and soft
operation research methods that can complement each other. In principle other
methods than the seven hard and seven soft methods presented can be used.
These methods have, however, been found to cover the actual demand for
methodology in the SP cases conducted so far. A not too high number of
different methods and techniques makes it possible for the facilitator and
analyst to acquire and master such a level of knowledge as regards the individual method that overall this will contribute to making the running of the
decision conference smoother.
VI. Ideally defining a number of adequate alternatives and afterwards assessing
these to determine the best choice should be based on a circularity between
scoping and assessment. As scoping and assessment are interdependent, the
identification of a number of relevant alternatives is in practice a processfinding based on the achievement among the participants of what has been
described as involved understanding (see below). Formally untimely scoping
can be expected to lead to suboptimising within the decision space. Such
suboptimisingleading possibly to a reasonably good decision may not later
on be recognised as such.
VII. The decision conference is undertaken as a way of accelerating and structuring learning given a specific complex plan problem. The decision conference in SP is seen as an interaction between a group of participants, the
application of formal decision analysis methods, and interactive IT, typically
large projected screens or smart boards supported by an analyst-operated PC.
When this interaction is working at its best the decision conferences made in
111
connection with the SP cases described above have been successful and have
obtained good evaluation questionnaire responses afterwards by the participants. Currently, however, no fixed recipe exists for organising and optimising SP decision conference sessions.
VIII. Applying SP was in theory seen as unfolding a subworld around the problem.
This makes sense as the subworld maintains a full perspective; at no stage in
the process should it turn into a solely calculative task that can be optimised.
Optimising can be provided in various respects typically as what-if questions behind specific model runs; however, the results obtained in this way are
still to be framed into a wider set of considerations. In practice the latter may as
was the case in the relocation case result in a wish to undertake a reexamination
of some of the issues debated and decided earlier. Even if such reexaminations
are not undertaken, the possibility to do so functions as a way to maintain a
practice-oriented comprehensive perspective.
Comparing SP to AHP very generally there seems to be less belief in SP than in
AHP that the generic decision making problem as described above can satisfactorily be based on one main methodology. The multitude of methods in the SP
toolbox simply indicates this viewpoint. In SP there also seems to be more explicit
concern about avoiding suboptimising and this is paid attention to by setting focus
on what has been described as the scoping-assessment circularity. In practice this
may not always lead to factual reconsideration of the choice alternatives once
again, but the awareness of the problem of untimely closure has in some of the
SP cases led to debates that have paid attention to considering going through the
alternatives once again; thereby the participants are made aware that it might be
relevant to reconsider some of the basic issues in the light of the way the process
has been conducted. Such reconsideration will help them to form an opinion
whether to accept (or to restrain from) ownership of the perception that the best
choice of alternative has been achieved.
Is SP a valid approach? Comparing it with AHP, which today is generally
accepted and tested through numerous cases (Saaty 2001; Bhushan and Rai 2004),
SP can borrow so-to-say validity from AHP as SP can be interpreted as a kind of
corrective and/or supplement to the problem handling in some of the steps in the
generic decision making problem as it was described above.
Practically SP has proved to be a flexible way of dealing with complex
problems; in this respect the underpinning theory was of value with regard to the
outline and design of the systemic process and toolbox. At the same time, however, the theory can be useful in another way, namely by dealing with validity and
potential and hereby more generally the relevance of strategic decision making
approaches such as AHP and SP in a wider context. In the following final section
of the book we will deal with this question of relevance by addressing the
possibilities and limitations of supporting complex strategic choices in a modern,
globalised and complex world.
112
113
Table 8.2 Problem types relating to the configuration of means and ends
Problem types Four different configurations and related approaches
Means/ends
Certain
Certain
A: Simplistic
Uncertain
Uncertain
Taleb has developed his Black Swan theory based on a background of a career
comprising both being a business investment trader and having a university career
as a professor in risk engineering. One can read his book as a major warning about
relying on long-term predictions that expose your decisions to unexpected events
such as extreme market moves. Specifically, he has set out a warning against
114
Table 8.3 Problem types relating to the configuration of means and ends
Talebs Fourth
Four quadrants ranging from extremely safe
Quadrant theory
to Black Swan domain
Domains
Simple payoffs
Complex payoffs
ending up in what he refers to as the Fourth Quadrant, see Table 8.3 above, where
radical change is possible and with severe consequences. Taleb communicates the
immanent nature of the environment of change by using the expressive labels
Mediocristan and Extremistan.
Taleb describes the quadrants in the following way:
First Quadrant. Simple binary payoffs in Mediocristan: forecasting is safe, life is easy,
models work, everyone should be happy. These situations are, unfortunately, more common in laboratories and games than in real life. We rarely observe these in payoffs in
economic decision making
Second Quadrant. Complex payoffs in Mediocristan: statistical methods may work
satisfactorily, though there are some risks. True, use of Mediocristan models may not be a
panacea There clearly are problems here, but these have been addressed extensively in
the literature
Third Quadrant. Simple payoffs in Extremistan: there is little harm in being wrong,
because the possibility of extreme events does not impact the payoffs. Dont worry too
much about Black Swans.
Fourth Quadrant, the Black Swan Domain. Complex payoffs in Extremistan: that is
where the problem resides; opportunities are present too. We need to avoid prediction of
remote payoffs, though not necessarily ordinary ones (Taleb 2010, pp. 363-365).
An important observation made by Taleb is that positive Black Swans are also a
possibility:
Actually, the Fourth Quadrant has two parts: exposure to positive or negative Black
Swans. I will focus here on the negative one The recommendation is to move from the
Fourth Quadrant into the third one. (Taleb 2010, p. 365).
115
116
117
in Sect. 8.1. Due, however, to the general nature of the SP framework as described
and validated in Sect. 8.2 there seems to be a potential for a broader use. In this
respect I would like to give attention to the recent development in business orientation and investment strategy known as social return of investment (SROI)
(Ellis 2010; Nicholls et al. 2010). A major recognition as regards SROI and related
ideas in what is generally known as corporate social responsibility (CSR), is that
these business trends will grow in importance to become major drivers in business
and management innovation (Ellis 2010). Specifically, Ellis sees the individual
organisations and companies as part of an ongoing trend seeking to merge their
core activities with CSR activities hereby creating value to both company and
society (Ibid., p. 157). Addressing this as Strategic CSR Ellis perceives this as a
new important business platform that makes it possible to set focus upon necessary
business innovation as well as upon wider sustainability. The latter is seen as both
a societal concern and a concern of the individual organisation and company.
With emphasis in this context on SROI as a way of determining the overall
effectiveness of Strategic CSR action, it can be noted that quite a number of
similarities exist between the ideas expressed in SP and those forwarded by SROI.
One main example is the necessity to address both monetised and non-monetised
issues; another one is the importance of including individual stakeholder views and
paying attention to the whole set of stakeholders involved in a particular strategic
action. In many contexts when addressing particular complex strategic choices on
the basis of an adopted CSR strategic company orientation, SP may be of interest
for decision-makers and analysts. The demo-case in Chap. 7 demonstrates how
management and staff can be dealt with as two stakeholder groups with different
interests in the relocation decision.
As concerns the challenge of strategic decision making as regards uncertainty in
what may collectively be addressed as the general decision environment, uncertainty is not expected to diminish in coming years. The challenge will consist of
strategic choices relating to what Nassim Taleb sees as Mediocristan as well as
Extremistan. No doubt especially the conditions with radical uncertainty characterising Extremistan call for special attention as concerns the formulation of the
strategic decision alternatives and the specific scenarios they are seen to be
embedded in.
The financial crisis, which is not yet over at the time of writing, has hit companies and organisations in ways that should make it relevant for them to generally
view their long-term planning and strategic choices in a complexity context. Doing
this could well pave the way for them to address more specifically the type of
decision aiding they really need to confront a world and a future characterised by
continuously rising uncertainty, where the financial crisis in a greater perspective
is but one of a number of possible precipitating causes. Even without any
knowledge of the financial crisis in 2008 and its continued reverberations, Martin
(2007) gives a compelling view of the hybrid of new and uncertain developments
and challenges across many societal fields. These are set out in both a short time
and a long time perspective in his thought-provoking book The Meaning of the
21st Century: A Vital Blueprint for Ensuring Our Future (Ibid.). One could narrow
118
Martins vision of the future down to the need and possibility of making better
complex strategic choices being the main concern behind the formulation of an SP
approach. In this respect I find in Martins outline of upcoming strategic decision
making challenges on many corporate and societal levels a thorough argumentation for considering more explicitly what was earlier described as the wider, nonmonetary performance of decision alternatives to be put in sufficient balance with
their core, monetary performance. Such an endeavour is behind both the theory
and practice chapters of the SP framework in the previous chapters.
One of the stepping stones towards the process and the methodology proposed
with the SP framework was the cognitive billboard, see Tables 5.1 and 5.2; another
was the determination of the seven hard and the seven soft operational research
methods contained in Table 5.5 and afterwards referred to as the toolbox. Both the
billboard and the toolbox in combination with the systemic learning approach
adapted (referred to as the subworld unfolding) have in their specific contexts of
applications so farrefer back to the case descriptions in Chaps. 7 and 8served
as ways of enhancing the creativity of decision-makers when addressing upcoming
strategic choice situations. With strategic cognition (SC) as a growing research
field per se and the SC research indicating and collecting evidence that cognitive
decision aids can improve decision quality (Narayanan et al. 2011, p. 341) there
seems to be some potential in continuing to develop SP and similar approaches.
Needless to say, both the cognitive billboard and the toolbox of SP will face
modifications in this respect when they are tested in new fields of application.
Hopefully the formulation of SP as set out in this book can inspire readers to
participate in such further development.
Main points and findings of this chapter
Systemic planning (SP) has been validated among other things by comparing it
to an approach to strategic decision making based solely on the multi-criteria
method analytic hierarchy process (AHP). It has been argued that SP has some
advantage because of its more explicit concern of qualitative information and its
use of soft and hard methodology in combination.
A number of SP application cases indicate that participants in decision conferenceswhich are really the integrating core activities of strategic decision
making in accordance with SPfeel comfortable with approaching complex
strategic choices by using SP. At the end of each decision conference (DC) that
has been conducted by the Decision Modelling Group at DTU Transport an
evaluation questionnaire has been distributed to the participants and returned
with generally very positive responses. In a theory context these responses, also
containing constructive suggestions made use of in various ways, show as a
relatively stable overall finding that in the course of SP activities and the concurrent deliberations of the DC participants, detached understanding and
deciding seems to recede and be replaced by involved understanding and
deciding about the complex problem dealt with.
119
References
Andersen M, Petersen NB (2006) Bedmmelse af omfartsveje i Allerd. M.Sc. thesis, Centre for
Traffic and Transport (CTT), Technical University of Denmark (DTU)
Barfod MB (2010) Structuring and appraising large and complex decision problems using
MCDA. In: Paper presented at the 24th European conference on operational research,
Portugal, July 2010
Barfod MB, Jensen AV, Leleur S (2008) STMA beslutnings-analyse: metoder, proces og
software. In: Carlsson CM (ed) Hllbart transportsystem fr inre och yttre attraktionskraft, EU
interreg project about strategic transport management in the resund region (STM) MAH
Malm, final STM report by Decision Modelling Group, DTU Transport. Technical
University of Denmark (DTU), Denmark
Beslutningsmodelgruppen (2009) Principper og metoder til brug for prioritering i forbindelse
med udmntning af pulje til mere cykeltrafik, DMG Danmarks Tekniske Universitet,
Denmark, Sept 2009
Beslutningsmodelgruppen (2010) Principper og metoder til brug for prioritering i forbindelse
med udmntning af ITS-puljen, DMG DTU Transport. Danmarks Tekniske Universitet,
Denmark, Dec 2010
Bhushan N, Rai K (2004) Strategic decision making: applying the analytic hierarchy process,
Springer series in decision engineering
Drewes Nielsen L, Homann Jespersen P (2003) The use of action research methods in scenario
construction. Seville workshop, Institute for Prospective TechnologiesEU Joint Research
Centres (IPTS), May 2003
DTU (2007) Dokumentationsmodel for TGB Vurderingsmodel (Trafikplan for Grnland:
Beslutningsredskab), Center for Trafik og Transport (CTT). Danmarks Tekniske Universitet,
Denmark
Ellis T (2010) The new pioneers: sustainable business success through social innovation and
social entrepreneurship. Wiley, New York
Hiselius LW, Barfod MB, Leleur S, Jeppesen, SL, Jensen AV, Hjalte K (2010) Helhetsorienterad
utvrdering av kollektivtrafiktgrder, Bulletin 246, Institutionen fr Teknik och Samhlle,
LTH, Lunds Universitet
120
Larsen, LA, Skougaard, BZ (2010) Vurdering af alternativer for en fast forbindelse HelsingrHelsingborg. M.Sc. thesis, DTU Transport. Danmarks Tekniske Universitet, Denmark
Leleur S (2000) Road infrastructure planninga decision-oriented approach, 2nd edn.
Polyteknisk Forlag Lyngby, Denmark
Leleur S (2008a) Systems science and complexity: some proposals for future development, Sys
Res Behav Sci 25(1):6779
Leleur S, Kronbak J, Rehfeld C (2000) The resund fixed link: evaluation issues and
development of new methodology, TRANS-TALK, ICCR, Vienna, EU 5th framework
programme, thematic network meeting, Brussels, Nov 2000
Leleur S, Holvad T, Jensen AV, Salling KB (2004a) Development of the CLG-DSS evaluation
model, midterm report presenting the CTT contribution to task 9 evaluation methodology,
Centre for Logistics and Goods Transport (CLG) at the Centre for Traffic and Transport
(CTT). Technical University of Denmark, Denmark
Leleur S, Jensen AV, Salling KB (2004b) COSIMASoftware manual, version June 2004,
Centre for Traffic and Transport (CTT). Technical University of Denmark, Denmark
Leleur S, Jensen AV, Salling KB (2004c) Modelling decision support and uncertainty for large
transport infrastructure projects: the CLG-DSS model of the resund fixed link. In: Antunes
CH, Dias LC (eds) Proceedings of the 15th mini-EURO conference in managing uncertainty
in decision support modelsMUDSM 2004, Coimbra, Portugal. INESC-Coimbra, Sept 2004
Leleur S, Petersen NB, Barfod MB (2007) The COSIMA approach to transport decision making:
combining cost-benefit and multi-criteria analysis for comprehensive project appraisal. In:
Kim K-S (Ed) Proceedings of the Korean Development Institute & World Bank Conference,
Seoul, May 2007. Improving public investment management for large-scale government
projects: focusing on the feasibility studies, Publisher KDI, pp 100122
Leleur S, Larsen AL, Skougaard BZ (2010) Strategic transport decision making: the SIMDEC
approach based on risk simulation and multi-criteria analysis, EUROSIS. In: Proceedings of
the Asian simulation technology conference ASTEC 2010 conference, Shanghai, March 2010
Martin J (2007) The meaning of the 21st century: a vital blueprint for ensuring our future.
Riverhead Books, Penquin Books, New York
Midgley G (2000) Systemic intervention: philosophy, methodology and practice. Kluwer
Academic, Kluwer
Mingers J, Gill A (1997) Multimethodology. Wiley, New York
Mitchell SD (2004) Why integrative pluralism? Emerg Complex Organ 6(12):8191
Mortensen, M, Andersen, MB (2007) Sammenfattende vurdering af luftfartsalternativer ved
Nuukmed srligt fokus p multi-kriterie analyse. M.Sc. thesis, Center for Trafik og
Transport (CTT). Danmarks Tekniske Universitet, Denmark
Narayanan VK, Zane LJ, Kemmerer B (2011) The cognitive perspective in strategy: an
integrative review. J Manag 37(1): 305351 (SAGE journals)
Nicholls J, Lawlor E, Neitzert E, Goodspeed T (2010) A guide to social return on investment, the
SROI network in association with the Scottish Government
Priemus H, Flyvbjerg B, van Wee B (2008) Decision-making on mega-projects: cost-benefit
analysis, planning and innovation. Edward Elgar Publishing, Cheltenham
Rumsfeld D (2002) http://en.wikipedia.org/wiki/There_are_known_knowns. Accessed 1 July
2011
Rnnest AK, Ohm A, Leleur S (1997) The resund fixed linkthe conflicts and the players,
WP7 case report, TEN-ASSESS research project, ICCR, Vienna, EU 4th framework
programme
Saaty TL (2001). Decision making for leaders: the analytic hierarchy process for decisions in a
complex world. RWS Publications, Pittsburgh
Salling KB (2008) Assessment of transport projects: risk analysis and decision support. Ph.D.
Dissertation, Department of Transport, Technical University of Denmark, Denmark
Salling KB, Leleur S (2009) Modelling of transport project uncertainties: risk assessment and
scenario analysis. In: Proceedings of the modelling and applied simulation09 conference.
Puerto de la Cruz, Tenerife
References
121
Taleb N (2010) The black swan: the impact of the highly improbable. Random House, New York
TetraPlan (2008) Traffic flow: scenario, traffic forecast and analysis of traffic on the TransEuropean Transport Network (TEN-T), TEN-CONNECT study, Final report for the EU
Commission, Dec 2008
Trafikministeriet (2003) Manual for samfundskonomisk analyse, juni 2003, Copenhagen
Tsamboulas D, Beuthe M, Grant-Muller S, Leleur S, Nellthorp J, Panou K, Pearman A, Rehfeld C
(1998) Innovations in decision analysis, deliverable D10, EUNET research project, EU 4th
framework programme
Appendix A
COSIMA
Overview
This appendix demonstrates the COSIMA approach that is one of the hard methods
in the seven hard and seven soft (2 9 7) methodologies in the SP toolbox, see
Table 5.5. The idea behind the composite model for assessment (COSIMA) is to
link cost-benefit analysis (CBA) with multi-criteria analysis (MCA). Thus CBA
provides an absolute, general assessment related to each alternatives core performance, whereas MCA provides a relative, context-dependent assessment related to each alternatives wider performance as set against the other alternatives
under examination.
COSIMA can be adapted to very different types of selection problems, where
criteria are available both in monetary and non-monetary terms. The Decision
Modelling Group (DMG) at DTU Transport at the Technical University of
Denmark has developed software that makes it possible to customise evaluation
models for a specific study.
Appendix A is an updated version of (Leleur et al. 2007) supplemented with
technical notes based on DMG (2010).
123
124
Appendix A: COSIMA
Many countries either use or have used CBA for transport decision making.
However, this method includes only impacts that can be valued monetarily. The
fundamental idea behind composite modelling assessment (COSIMA) is to extend
conventional cost-benefit analysis (CBA) into a more comprehensive type of
analysisas often demanded by decision-makers (DM)by including missing
decision criteria of relevance for the actual appraisal task. The missing criteria
often address issues that have been difficult to assess by the conventional CBA but
which hold a potential of improving the actual decision support from the appraisal
if treated properly. This is the purpose of COSIMA, where the added criteria will
be referred to as the multi-criteria analysis (MCA) part of the COSIMA analysis.
The COSIMA method will be described in detail below and thereafter
demonstrated on two cases to show the features of the method and the possibilities
it offers. The two cases concern an examination of alternatives for a by-pass road
around the Danish town of Hng and for a new ring road in conjunction with a new
residential area in the town of Allerd. Finally, a discussion of COSIMA compared
to the CBA and MCA methods is undertaken, and the methods are compared with
focus on the following three issues (Andersen and Petersen 2006):
Comprehensiveness: As previously mentioned, not all effects can be treated in the
same way. This creates a challenge for the methods as they have to be able to
include all important effects. The comprehensiveness issue describes how well the
methods succeed in doing this.
Effectiveness: The effectiveness issue describes how easily a final choice can be
made on the basis of the result of the analysis. An analysis must be able to consider
all important effects but also to make the results usable for the decision-makers.
Transparency: Transparency is important with regard to transportation decision
methods because the choices made by the decision-makers must be understood and
accepted by the public. This does not necessarily mean that there must be complete
consensus about the choices made but that a decision which is difficult to
understand for the public is more likely to face opposition and thereby possibly
entail expensive delays or rejection.
Appendix A: COSIMA
125
126
Appendix A: COSIMA
Fig. A.1 Non-linear and linear value functions. The latter are assumed in SMARTER
set against the other alternatives in the examination. For the MCA criteria a setspecific (local) scale assigns the value 0 to the worst performing alternative and the
value 100 to the best performing alternative. The rest of the alternatives are then
rated by relating them to these alternatives (Belton and Stewart 2002,
pp. 121122). With application of point scales and/or AHP, dependence on the
actual set of alternatives is introduced. In practice this can be dealt with in a
satisfactory way within the model calibration and as part of the COSIMA decision
conference described later. Hereafter three stages I, II and III remain:
Stage I. With the CBA and MCA effects specified, the so-called anchoring
part of the COSIMA model formulation can take place, which concerns
determining the importance of the MCA effects against the CBA effects, i.e. the
overall MCA versus CBA trade-off, and for the MCA effects among each other,
i.e. the determination of MCA criteria weights. With regard to the latter, several
MCA techniques can be made use of: direct weights, pairwise comparisons, swing
weights, etc. (Ibid., pp. 134143, pp. 157159). To ease the assignment of criteria
weights for the MCA effects that can represent the actual DM preferences, the rank
order centroid (ROC) or the more recent rank order distribution (ROD) weighting
technique are also applicable. The determination of rank order weights is based on
the assumption that weights assigned by the decision-makersby simply ranking
the actual MCA criteriacan be derived by using specified probability density
functions (Roberts and Goodwin 2002; Goodwin and Wright 2010, pp. 6366).
The choice of relevant MCA effects and the assignment of weights to these effects
will usually be determined during a number of decision conference sessions, where
both DM and decision analysts take part, see the description later.
Stage II. After the MCA effects and their assigned weights have been agreed
upon, COSIMA can be run. As previously mentioned, COSIMA includes the MCA
effects or criteria along with those usually treated in a CBA, thereby calculating a
total value (TV) in monetary units for alternative Ak obtained by spending the
investment cost Ck:
TVAk = CBAAk + MCAAk
A:1
Appendix A: COSIMA
127
VCBA Xik a
w j VMCA Xjk
Ck
Ck
i1
j1
A:2
with:
VCBA(Xik): Value in monetary units for CBA effect i for alternative k for
altogether I CBA effects. During model calibration it is kept fixed as bik.
VMCA(Xjk): Value function rating for MCA criterion j for alternative k for
altogether J MCA criteria. During model calibration it is transformed into a
monetary value bjk.
a: Calibration factor that expresses the specific models trade-off between the
CBA and the MCA part. It should be observed that the CBA calculation remains
unchanged, but that different values of a will change the influence of the MCA
on the TRR value. The value of a = a(MCA%) is set by specifying
MCA% = 100Rj(Bj)/[(Ri(Bi) + Rj(Bj)], where Aj denominates a subset of the
k = 1..K alternatives Ak (with this subset selected for calibration, see the
Technical Notes accompanying Appendix A below) and Bi = Rj[K(bij) and Bj =
Rj[K(bjj) enter as the value elements for the individual effects i and criteria
j summed over the j alternatives; thus Ri(Bi) and Rj(Bj) concern row summations
over the I CBA effects and the J MCA criteria and Bi and Bj the results of bik and
bjk column summations over the alternatives, where some if not all are selected
for the model calibration.
w(j): Weight that expresses the influence of criterion j.
The general COSIMA principles are presented by (A.1) and (A.2). It can be
seen that with sufficient information about the MCA part, (A.2) can be specified
into a CBA-like calculation. This will be the situation when, for example,
a conventional CBA is carried out, and it is afterwards (1) supplemented with
some extra criteria which can be specified fully by impact models that (2) can
determine net effects which (3) can be given satisfactory unit prices similar to the
assessment in the CBA part. Most often, however, this will not be possible,
because usually the MCA part will be less known than the CBA part. In fact the
purpose of COSIMA is to handle such a situation. In modelling terms, this can be
done by the determination of appropriate values for a and w(j) for the J MCA
criteria and by the determination of appropriate value function ratings VMCA(Xjk).
128
Appendix A: COSIMA
The latter supplement the determination of VCBA(Xik) that can be derived from a
CBA manual relevant for the actual assessment case.
Stage III. At this stage COSIMA is run for all the project alternatives. Then the
model inputs and the related assessment questions are addressed on the basis of
the results obtained and the assumptions behind them, and a new exchange
with the decision-makers (DM) is carried out with two principal possibilities
available now. The study may simply end here if the DM are confident about the
model outcome, or the DM may want to go back into the process and re-address
some of the previous model settings to shed light on some issues that have caught
their attention.
One important characteristic of the COSIMA approach is that the model is more
or less customised to the specific appraisal case. It should be observed, however,
that the assessment result produced is given as total rate of return (TRR) values
stemming from an objective CBA part and a more subjective MCA part, where
the CBA part represents a result in its own right, i.e. without the MCA add-on
which actually only provides the DM with some extra discriminatory
information. The CBA-like way applied in COSIMA to present both the CBA
and MCA assessment information may appeal to decision-makers who want the
possibility of refined analysis using all the available information but kept in a
simple and straightforward way. To illustrate this, the decision-makers may wish
to use only the benefit-cost rate (B/C) part of the TRR value from a base case
scenario without any further analysisthis would in fact be a conventional
analysisor they may like to inspect some or all the TRR values and their
composition as produced in a number of what-if scenarios.
Figure A.2 shows the locations of the two case examples presented in this
appendix. The first case concerns the assessment of seven by-pass alternatives for
relieving the town of Hng in a rural area in western Zealand of through traffic.
The second case deals with four alternatives for a new ring road around the town
of Allerd situated in an urban area north of Copenhagen in northern Zealand.
Both case examples are representative of a typical appraisal task in
infrastructure investment planning: a mix of hard (CBA) and soft (MCA) issues
that are relevant for the decision-makers to take into consideration when selecting
the most attractive alternative.
Appendix A: COSIMA
129
noise, air pollution and severance and perceived risk) and the following three
MCA criteria: network accessibility, urban planning and landscape.
By use of the methodology from the Danish Road Directorate, first-year
benefits (FYB) have been calculated for the seven alternatives (Leleur 2000). This
information has been put together with point scores for the three MCA criteria,
where the point scores are determined by thorough examination of the alternatives
based on a rating protocol. Hereby the project-effect matrix in Table A.1 has been
formulated (Steffensen and Testmann 2000).
At this stage the next step is anchoring the MCA criteria. The three MCA
effects are assigned a value describing their performance on a scale from -5 to +5,
where +5 is best.
The scores are then translated into ratings between 0 and 100, see Table A.2,
using a linear, local value function (Belton and Stewart 2002, pp. 121122).
The MCA impacts must then be assigned weights to be used in the COSIMA
analysis. As the ROD weights are used, the effects need only be ranked by the DM
with the weights themselves being predetermined according to the ROD weight
principles (Roberts and Goodwin 2002). Any ranking could have been used, but it
is assumed that the decision-makers have agreed on the ranking in Table A.3 and
hereby indirectly on the shown weights.
The ratings and weights of the MCA effects are entered into the COSIMA
software along with the normal CBA input for each alternative. It should be noted
that traffic forecasts are also included and that the COSIMA results are based on
discounting the values for a 30-year service-period back to the opening year.
130
Appendix A: COSIMA
Table A.1 The three MCA impacts together with the investment cost and first-year benefits
(FYB) from the CBA methodology
Alternatives
Cost in m DKK
FYB in k DKK
Network
Urban
Landscape
accessibility
planning
Alternative
Alternative
Alternative
Alternative
Alternative
Alternative
Alternative
1
2
3
4
5
6
7
16.7
15.3
16.9
18.0
17.2
20.8
19.9
1,310
790
350
2,900
2,460
1,790
2,330
+2
+1
-1
+4
+3
-3
+3
+1
+1
+3
-2
+2
+4
+1
-2
+4
+2
+2
+3
0
+3
Table A.2 The three MCA impacts rated using the point scale method
Alternatives
Cost in
FYB in k DKK
Network
Urban
m DKK
accessibility
planning
Landscape
Alternative
Alternative
Alternative
Alternative
Alternative
Alternative
Alternative
0
100
67
67
83
33
83
1
2
3
4
5
6
7
16.7
15.3
16.9
18.0
17.2
20.8
19.9
1,310
790
350
2,900
2,460
1,790
2,330
71
57
29
100
86
0
86
50
50
83
0
67
100
50
Effect ranking
w(j)
1. Network accessibility
2. Urban planning
3. Landscape
0.52
0.33
0.15
Afterwards a is determined so the CBA and MCA parts of the analysis are traded
off in a way found suitable by the decision-makers. In this case an MCA% = 50
split is used. The programme then calibrates the model so that the MCA fraction of
the total benefits equals the split specified by the user. The case example 1 results
are shown in Table A.4.
Figure A.3 indicates how the total rate of return (TRR) values consist of both
the CBA and MCA impact contributions for all the alternatives.
It is clearly seen that alternative 5 scores higher than the other alternatives due
to a reasonably good performance with regard to the B/C-rate and the MCA
criteria. The MCA criteria are assessed by trade-off implied unit prices, set against
the costs of the individual alternative and added to the CBA part; in principle only
the latter is a monetary return of the investment (the costs of the alternative) and
the MCA rate solely a value expression of further information to make it possible
to discriminate between the alternatives.
Appendix A: COSIMA
131
Urban
planning
Landscape
TRR values
Alternative
Alternative
Alternative
Alternative
Alternative
Alternative
Alternative
0.58
0.64
0.96
0.00
0.76
0.94
0.49
0.00
0.60
0.36
0.34
0.45
0.15
0.39
3.59
3.51
2.29
5.49
5.81
2.91
4.71
1
2
3
4
5
6
7
1.66
1.09
0.44
3.40
3.03
1.82
2.47
1.35
1.18
0.53
1.75
1.57
0.00
1.36
132
Appendix A: COSIMA
10
9
Alternative 1
Alternative 2
Alternative 3
Alternative 4
Alternative 5
Alternative 6
Alternative 7
TRR
6
5
4
3
2
1
0
0%
10%
20%
30%
40%
50%
60%
70%
MCA %
alternatives to consider, and it can also help sort out the lowest scoring alternatives
in the entire range.
Appendix A: COSIMA
Table A.5 Ratings assigned
to the three MCA effects by
using the AHP method
133
Alternatives
Cost in
m DKK
Accessibility
Local
land use
Regional
network
Alternative
Alternative
Alternative
Alternative
19.2
54.9
48.6
194.9
0
46
19
100
0
100
7
100
0
23
3
100
1
2
3
4
Effect ranking
w(j)
1. Accessibility
2. Local land use
3. Regional network
0.52
0.33
0.15
Alternatives
B/C
Alternative
Alternative
Alternative
Alternative
2.03
1.05
1.52
0.50
0.00
0.10
0.01
0.13
1
2
3
4
0.00
0.66
0.31
0.43
0.00
0.89
0.07
0.27
2.03
2.70
1.91
1.33
see Table A.5. Note that alternative 1 is rated as the lowest for all the MCA effects,
while alternative 4 is rated best. It can be noticed that the AHP assessment depends
on the actual alternatives; this may be more pronounced than it could be observed
in the previous case example 1 applying the formulated point scale. Although such
dependencies can be seen as less attractive from a theoretical viewpoint, this
method aspect is due to the more subjective nature of MCA, see the earlier
discussion of the CBA and MCA part of COSIMA. The various theoretical and
practical issues relating to the application of AHP are treated thoroughly by
(Belton and Stewart 2002).
The MCA effects are ranked according to importance and weighted using the
ROD weights as seen in Table A.6. As in case example 1, any ranking could have
been used.
The CBA is based on the Manual for Socio-Economic Analysis
(Trafikministeriet 2003). The TRR values for COSIMA with 50% weight on the
MCA are seen in Table A.7.
Table A.7 and Fig. A.5 indicate how the total TRR values are made up by the
CBA and MCA effects.
Alternative 1 only has a contribution from the CBA effects, but it still scores
higher than alternatives 3 and 4. Due to a large contribution from the MCA effects,
alternative 2 has a better TRR than the other alternatives.
Figure A.6 shows that alternatives 1 and 2 are possible contenders to
finally being selected as the most attractive choice. For a low percentage
134
Appendix A: COSIMA
Alternative 1
Alternative 2
Alternative 3
Alternative 4
TRR
6
5
4
3
2
1
0
0%
10%
20%
30%
40%
50%
60%
70%
80%
MCA %
assigned to the MCA, alternative 1 scores highest, while alternative 2 is the most
attractive alternative for a relatively high percentage assigned to the MCA.
Alternative 4 is the least attractive one for all trade-off percentages except for the
MCA% equal to 70 or 80.
The results are presented to the decision-makers, who are then able to make an
informed decision based on both the usual CBA and the COSIMA approach with
varying percentages assigned to the trade-off between CBA and MCA.
Appendix A: COSIMA
135
The MCA% values that ought to be applied in the concrete appraisal task
depend on the decision-makers and their interpretation of which amount and type
of factors or criteria they want to influence their decision. Practical experience so
far points to MCA% values in the range of 1050. Furthermore, it seems that high
MCA% values are most likely to be adopted when appraising larger and more
complex transport infrastructure projects.
136
Appendix A: COSIMA
Appendix A: COSIMA
137
example that was finished in the middle of 2007 is a COSIMA version (TGB) for
the planning of new airfields in Greenland (DTU 2007). In addition to the basic
features treated in this appendix, COSIMA-TGB contains features for the handling
of the complex data input from flow modelling and from accessibility modelling,
which is one of the MCA criteria (Kronbak 1998). Furthermore, the model is set up
to treat various taxation issues and growth scenarios, and it is also endowed with
features for quantitative risk analysis (QRA) based on Monte Carlo simulation
(MCS) (Vose 2002; Leleur et al. 2004a, b; Salling 2008).
One basic issue to be addressed when setting up a new COSIMA model
application concerns its calibration. In the programme it is possible to select the
alternatives that are to be used for the calibration of the unit prices. Usually, all
alternatives that are thought to be serious contenders as a final choice are used for
the calibration. Alternatively, the project alternative with the highest B/C-rate can
be used for the calibration. The two calibration methods can give slightly varying
results, but which of the two methods to apply may depend on whether the project
with the highest B/C-rate is seen as being challenged by all others or, for example,
a minor or major group is seen as more or less equal candidates.
When calibrating, the user must make sure that all MCA effects are taken into
account. Therefore, if the alternative with the highest B/C-rate does not have a
rating above 0 in all effects, this alternative alone cannot be used for the calibration
and more alternatives must be used. In the case examples 1 and 2, described to
demonstrate the COSIMA approach, it is not possible to calibrate from the
alternative with the highest B/C-rate, as these alternatives do not have
contributions from some or all MCA effects. Therefore, all alternatives are used
for the calibration as all alternatives are thought to be possible contenders for a
final choice. When relating to the appraisal of large transport infrastructure
investments, COSIMA can also be calibrated to accommodate computable general
equilibrium (CGE) analysis, see (Leleur and Holvad 2004) for an analysis of the
resund Fixed Link.
138
Appendix A: COSIMA
The MCA as generic method can include all possible effects as it is not
necessary to be able to assign monetary values to the MCA effects. Therefore, DM
wanting one method to include all effects could be tempted to use the pure MCA
approach. However, the composite method COSIMA can also include all effects,
but as described they will be split into effects that can and cannot in principle be
assigned a monetary value. Furthermore, when using the COSIMA method, the
result is a total rate of return which shows how the benefits are made up of the
CBA part and the different MCA effects. This means that the COSIMA method
provides the DM with CBA based information about socio-economic viability,
which cannot be provided in the MCA. Furthermore, MCA is a comparative tool
only and therefore it does not tell the decision-makers if the benefits of the project
at hand exceed the costs. COSIMA thus combines some merits, as it gives
information about socio-economic viability while at the same time being able to
include all effects like the MCA.
It is not correct to refer to the MCA approachalthough it is only done
genericallyas being one method as many methods are available (Tsamboulas
et al. 1998; Belton and Stewart 2002). In this context the MCA method therefore
refers only to some common traits across a number of individual MCA methods.
However, to ease comparison issues, one could see the SMART approach
developed by von Winterfeldt and Edwards (1986) as an exemplar MCA
methodology that is perceived as a recommendable standard approach (Goodwin
and Wright 2010, pp. 3156). Very briefly one could see the COSIMA approach
and the SMART approach as appraisal methodologies that follow opposite
strategies: where SMART translates what in this context has been described as the
CBA impacts into MCA-like value functions, COSIMA translates the MCA
impacts and their associated value functions into a CBA-like comprehensive type
of analysis, see Eqs. A.1 and A.2. Below CBA, MCA and COSIMA will be
compared further as concerns efficiency and transparency.
Efficiency describes how easily a final decision can be made on the basis of the
results of the analysis. The CBA itself gives a very clear answer as to which
alternative is the most attractive as it presents a single point estimate for each
alternative. However, as not all effects can be included in this analysis, it is
necessary to take these into account separately, and the decision will easily
become more complex.
Similarly, the MCA and COSIMA approaches per se give a final, single
appraisal result for each alternative based on considering all the effects. Therefore,
the final decision on the basis of these two methods is very straightforward,
although it is necessary to scrutinise the applied MCA scores for the different
alternatives and the appropriate CBA/MCA split. With regard to COSIMA it
should be noted that the MCA part is context-dependent by setting focus on the
relative performance of an alternative as compared to the worst of alternatives
under a specific criterion. Therefore the set of alternatives needs to be made up by
alternatives where each of these under a specific viewpoint represents a sound
candidate for being selected for implementation. At the same time it must be
required that each criterion in the criteria set should help to discriminate in an
Appendix A: COSIMA
139
adequate way between the alternatives. If this is not the case (in AHP leading to
very similar scores across the alternatives) this should lead to the exclusion of that
criterion or to the inclusion of an alternative that changes this situation but at the
same time is seen as a sound alternative worthwhile to consider.
Transparency is important in transport decision making to prevent, for example,
public opposition to the decisions taken. The traditional CBA has a high degree of
transparency as the analysis in theory can be redone by anyone. The assigned unit
prices are predetermined ensuring that the project will be appraised in the same
way no matter who carries out the analysis.
Both the MCA and COSIMA assume the presence of the DM during the
calculation process, and the decision-makers are required to make choices
regarding which effects to include and how to rate and rank them. This entails the
need for thorough discussions among the decision-makers and in order to ensure
transparency of the choices taken the reasons behind them must be presented. It is
considered an advantage that the decision-makers are not asked to determine
specific weights directly and that a COSIMA decision conference can be run in a
more or less non-technical way.
The procedures in the MCA method can be very straightforward and easy to
follow for the decision-makers. The COSIMA method may appear to be slightly
more demanding to the decision-makers, for which reason its principles should be
explained, for example, by use of a demo-case at the beginning of the first decision
conference meeting before the described DM/CF questionanswer session begins.
However, the results of the COSIMA method, the TRR values, could be easier to
understand than MCA results as the CBA method is presumed to be well known to
the DM.
The involvement of decision-makers in the entire process demands
transparency but is a great advantage as it makes it possible to adjust the
analysis to local conditions and the specific requirements of the decision-makers.
What is weighted highest in a transportation project can differ from project to
project and between decision-makers. Therefore, DM in one planning context and
location, for example, might assess a problem differently from decision-makers in
another context. Both the MCA and COSIMA approaches allow the users to
influence the setting of weights of the effects according to their wishes, for
example by determining suitable rankings.
Summing up the COSIMA method combines the comprehensiveness of the
MCA with the information given in the CBA about socio-economic viability.
Furthermore, the method allows the decision-makers to include the MCA effects of
their choice and rank and weight them according to their perceptions of importance.
140
Appendix A: COSIMA
Appendix A: COSIMA
141
1:
2:
3:
4:
5:
6:
7:
8:
9:
In Stage 7 the least important criterion from Stage 6 is assigned the value 10.
The second-least important criterion is given a value that indicates its value
relatively to the least important criterion and so on until all the criteria are assigned
values. These are then normalised into weights summing up to 1.
In SMART, the ratings of the alternatives in Stage 5 are assigned directly by
using appropriate, natural scales of the criteria. For instance, when assessing the
criterion cost for the choice between different road layouts, a natural scale
would be a range between the most expensive and the cheapest road layout. In
order to keep the weighting of the criteria and the rating of the alternatives as
separate as possible, the different scales of the criteria need to be converted into a
common internal scale. In SMART, this is done mathematically by the decisionmakers by means of a value function. The simplest and most widely used value
function is a scale going from 0 to 100.
SMART Exploiting Ranks
The assessment of value functions and swing weights in SMART can sometimes
be a difficult task, and decision-makers may not always feel confident about it.
Because of this, Edwards and Barron have suggested a simplified form of SMART
named SMARTER (Roberts and Goodwin 2002). Using the SMARTER technique
the decision-makers place the criteria into an importance order, for example
Criterion 1 is more important than Criterion 2, which is more important than
Criterion 3, which is more important Criterion 4 and so on leading to C1 C C2 C
C3 C C4 . Afterwards SMARTER assigns weights by using the rank order
distribution (ROD) method.
142
Appendix A: COSIMA
ROD is based on a weight approximation that assumes that valid weights can be
elicited by the ranking of criteria; this is very convenient in a decision conference
as ranking is more easily negotiated than are direct weights.
The approximated ROD weights for n = 2 to 10 have been found
mathematically. For further information about the underlying calculations, see
(Roberts and Goodwin 2002). Weight sets are indicated below for a number of
criteria between two and eight with rank order of criteria from left to right. Note
that Roberts and Goodwin originally indicate the weights with four decimals,
which may, however, indicate a kind of precision that is not realistic. They also
indicate weights for nine and ten criteria but as can be seen from the values below
more than eight criteria will mean that practically no discriminatory power is given
to the criteria ranked as numbers nine and ten. This information is relevant for the
participants in a decision conference when engaged in reducing an initially long
list of criteria, see the case examples in Sect. 8.1.
Two criteria
Three criteria
Four criteria
Five criteria
Six criteria
Seven criteria
Eight criteria
(0.69;
(0.52;
(0.42;
(0.34;
(0.30;
(0.26;
(0.23;
0.31)
0.33; 0.15)
0.30; 0.19; 0.09)
0.27; 0.20; 0.13; 0.06)
0.24; 0.19; 0.14; 0.09; 0.04)
0.22; 0.18; 0.14; 0.10; 0.07; 0.03)
0.20; 0.17; 0.14; 0.11; 0.08; 0.05; 0.02)
(1)
(2)
(3)
(4)
(5)
(6)
Appendix A: COSIMA
143
(7)
(8)
(9)
With two alternatives (or criteria) A and B compared, the DM are asked to
indicate the preference intensity. If, for example, A is preferred to B with strong
importance grade 5 is chosen (if B is preferred to A with the same intensity the
reciprocal value 1/5 is indicated). In this way a positive comparison matrix aij is
set out with the reciprocal property that aij = 1/aji and the diagonal-elements equal
to 1. Furthermore preference transitivity leads to aij = aik/ajk.
In the original approach by Saaty the PerronFrobenius theory is used to
determine the preference weights (relative performance indicators) for the
alternatives with respect to each criterion and on the next level for the criteria
with respect to the goal. Saaty found that the problem of determining these
preference weights can be formulated as an eigenvector problem (Hwang and
Yoon 1995). By multiplicating upwards in the hierarchy the overall relative
performance of each alternative with respect to the goal can then be determined.
Due to its intuitive appeal and the development of software such as Expert
Choice, AHP gained widespread use (Expert Choice Inc. 2004). This success is no
doubt due to the fact that decisions produced in an AHP session have generally
appeared to be in accordance with the preference-views expressed by the DM
participating in the AHP process (Vaidya and Kumar 2006).
It should be noted, however, that criticism has been expressed concentrating on
the following three issues (Olson et al. 1995; Lootsma 1999):
The 19 ratio scale applied to quantify the preference judgments
The application of the PerronFrobenius eigenvector method
The arithmetic mean aggregation rule
It can be noted that these criticisms have been dealt with by introducing a
multiplicative version of AHP called the REMBRANDT method, which is
based on (1) applying a difference-based scale judging the preference intensity for
the individual pairwise comparisons, (2) replacing the eigenvector method by
geometric mean aggregation (therebyin case of adding a new alternativerank
reversal of presently examined alternatives can be avoided) and (3) by replacing
the weighted arithmetic mean aggregation of scores with aggregation by the
product of relative scores lifted into the power of the criteria weights (Ibid.).
It has been concluded that both the ratio-based AHP and the difference-based
REMBRANDT methods are useful, practical multi-criteria decision making tools
(Barfod and Leleur 2011). The REMBRANDT method is explained as part of the
Technical Notes accompanying Appendix B below.
A thorough treatment of AHP for strategic decision making is given in (Saaty
2001; Bhushan and Rai 2004). As mentioned in Chap. 8 the latter reference has
144
Appendix A: COSIMA
Appendix A: COSIMA
Table A.8 Benefit-cost data
for the calculation example
145
Alternatives
A1
A2
A3
A4
B1..B4
C1..C4
BCR
110
70
1.57
160
80
2.00
165
120
1.38
120
65
1.85
Numerical Example
The COSIMA calculations can be illustrated with a simple calculation example
based on (Hiselius et al. 2010).
In the present alternative survey four alternatives A1, A2, A3 and A4 are
available. Using a national cost-benefit manual and its fixed unit price values the
total benefits are calculated: B1, B2, B3 and B4, which by dividing them with the
observed total expenditure C1 C2, C3 and C4 leads to benefit-cost rates (BCR) for
the four alternatives, see Table A.8.
If the DM agreeafter the content of the CBA is reviewedthat the decision
making is complete, a decision to choose A2 can be taken, since this alternative
has the highest BCR value = 2.
If the CBA is insufficient, new criteria can be added to evaluate them by a MCA
and finally perform a composite analysis according to the COSIMA principles. The
procedure is as follows: first a number of criteria are described, which in this
example leads to the criteria k1, k2, k3 and k4 that are determined in such a way
that an overlap with the components of the CBA is avoided.
Next, the four criteria are rated and weighted. The rating means that each
alternative for each criterion is assigned a value (score), which lies between 0 and
100. The value 0 is given to the alternative that is performing worst under the
given criterion and 100 to the alternative which is performing the best. The two
remaining alternatives will have values between 0 and 100. The approach is based
on pairwise comparison of all four alternatives under each of the four criteria k1,
k2, k3 and k4. For each of these criteria with four alternatives examined altogether
(4 9 3)/2 = 6 pairwise comparisons are needed. Based on the MCA method
REMBRANDT (see Appendix B) the following scores are obtained from a
transformation of the REMBRANDT results into a value function, see Table A.9.
Since the criteria are usually not assigned equal importance by the decisionmakers, the criteria are assigned the weights K1, K2, K3 and K4. This can be done
directly or by using the ranking criteria method ROD. The result, where the
weights are set directly and summarise to 1, is for this example: (K1, K2, K3,
K4) = (0.20; 0.55; 0.10; 0.15).
In the last part of the calculation the CBA and MCA are linked together, which
is done by decision-makers providing the MCA%. At a high MCA% the MCA will
dominate the final result, while a low MCA% means that it will be the CBA and
the BCR values that dominate.
146
Appendix A: COSIMA
Criteria/Alternative
A1
A2
A3
A4
k1
k2
k3
k4
25
0
0
100
100
75
26
68
0
60
100
35
45
100
35
0
The decision-makers are asked about the MCA% and they decide, for example,
to set this CBA/MCA trade-off parameter to 50%. Thus MCA and CBA count the
same in the overall analysis. Based on the choice of A2 with the highest BCR as
calibration basis, the MCA part should now count the same. Benefit value B2
was found to be 160 which means that the MCA part of A2 should also sum up to
160. Adding up the MCA-components of A2 using the scores in Table A.9 p1 can
be determined in the following manner, with p2, p3 and p4 expressed by p1 and the
criteria weights:
100 p1 75 p2 26 p3 68 p4 160 >
100
0:20
0:55
0:10
0:15
p1 75
p1 26
p1 68
p1 160
0:20
0:20
0:20
0:20
TRRA1
TRRA3 2:25
TRRA4 4:09
From this it is seen that A4 is the most attractive alternative. In Table A.10 the
results are shown in overview. The example is based on the use of cost-benefit
analysis (CBA) carried out by use of a national manual (NM) and a multi-criteria
analysis (MCA). CBA + NM produce a monetary result, which is validated from
socio-economic thinking and common use. MCA produces a result which is based
on preferences indicated in the decision process and the result is in principle only
valid from this point of view.
Appendix A: COSIMA
147
70
110
1.57
11
0
0
32
43
153
2.19
80
160
2.00
43
89
6
22
160
320
4.00
120
165
1.38
0
71
22
11
104
269
2.24
65
120
1.85
19
119
8
0
146
266
4.09
CBA +
CBA +
CBA +
MCA
MCA
MCA
MCA
MCA
CBA +
Unit
NM
NM
NM
m DKK
m DKK
eval m DKK
eval m DKK
eval m DKK
eval m DKK
eval m DKK
NM + MCA m DKK and eval m DKK
With the CBA benefits expressed in million DKK (m DKK) the MCA results
(MCA benefits) will also be expressed in million DKK, here to be expressed as
evaluation DKK (eval m DKK) to indicate their way of determination. With the
TRR based on both types of benefits (m DKK and eval m DKK) the TRR benefits
are therefore a mix of m DKK and eval m DKK. The total rate for an examined
alternative is found by adding the benefits from CBA and MCA and dividing them
by the costs in m DKK. Table A.10 presents the results of the numerical example.
The final result is determined by having A2 as the basis for the calibration and
the MCA% set to 50.
By means of the COSIMA calculations it is possible to base the decision of the
choice of alternative on the socio-economic BCR core contribution in combination
with its MCA-based wider performance. This combined result expresses the
overall attractiveness of a given alternative. As mentioned, the CBA result is valid
in the light of being provided by a socio-economic evaluation, while the MCA
result in principle only is valid on the basis of the conducted decision conference
and the actual deliberations that have taken place. A better background can be
obtained if a log book is worked out with user inputs and background comments.
This allows a second-opinion to be set out by inspecting these.
As stated, the values in Table A.10 are developed on the basis of the phrase
influence of the MCA must be 50%. How can this be interpreted further? The
basis for choosing among A1, A2, A3 and A4 is a CBA, which shows that the A2
due to the highest BCR (= 2.00) is the best choice. This BCR value for A2 is given
by C2 = 80 and B2 = 160. A balance between CBA and MCA must be arranged
so that the MCA criteria indirectly priced also contribute with 160, which has just
been illustrated in the calculation example. Keeping the scores of the alternatives
and the criteria weights unchanged this determines unique total rates (TRRs) for
all four alternatives, and A4 stands as the most attractive alternative. In brief the
COSIMA analysis replaces A2 by A4 as the most attractive choice.
Appendix B
SIMDEC
Overview
This appendix demonstrates the SIMDEC approach that is one of the hard methods
in the seven hard and seven soft (2 9 7) methodologies in the SP toolbox, see
Table 5.5. The idea behind risk simulation and multi-criteria analysis in combination for decision making (SIMDEC) is to incorporate risk analysis (RA) by
using Monte Carlo simulation (MCS) as one specific criterion within a multicriteria analysis (MCA). Absolute assessment of each examined alternatives core
performance is provided by RA as a probability-based interval result, whereas
MCA provides a relative, context-dependent assessment related to each alternatives wider performance as set against the other alternatives under examination.
SIMDEC can be adapted to very different types of selection problems, where
criteria are available both in monetary and non-monetary terms. The Decision
Modelling Group (DMG) at DTU Transport at the Technical University of
Denmark has developed software that makes it possible to customise evaluation
models for a specific study.
Appendix B is an updated version of (Leleur et al. 2010) supplemented with
technical notes based on DMG (2010).
149
150
Appendix B: SIMDEC
B.2.1
The FRA is carried out by using Monte Carlo simulation (Vose 2002; Salling
2008) on the results stemming from a conventional cost-benefit analysis, which is
assumed to be prescribed by and conducted in accordance with a manual that can
generally be accepted in the study context. For a large transport infrastructure
investment the impacts to be covered will consist of: construction and maintenance
costs, time savings, operation costs, accident savings, noise emissions, local air
pollution and climate effects based on change in CO2 emissions. With the
exception of construction costs and time savings these effects can be determined in
Appendix B: SIMDEC
151
The MCA is carried out by using the REMBRANDT technique (Olson et al.
1995; Olson 1996) based on pairwise comparisons for rating of the alternatives and
determination of the criteria weights. The REMBRANDT technique is recognised
as a both valid and practical framework (Lootsma 1999; Barfod et al. 2011; Barfod
and Leleur 2011).
In SIMDEC a set of relevant decision criteria for the decision problem at hand
is laid down. Generally such a set will consist of both monetary and non-monetary
criteria. As already indicated the monetary criteria are taken into account by using
a CBA, which again is used as an input to conducting a feasibility risk assessment
(FRA) of each of the alternatives. Specifically a certainty graph, CG(x), is
produced for each of the alternatives. For each alternative this graph represents its
FRA-performance (based on the monetary criteria and the conducted MCS). This
FRA-performance is added as a criterion to the formulated non-economic criteria
that typically represent strategic issues and impacts relating to the decision
problem. Thereby the total criteria set is established for the examination of the
decision problem. This set should be scrutinised to reduce possible overlapping
with regard to criteria definitions, while at the same time it should be ensured that
no valuable information for the decision making has been left out.
The multi-criteria analysis proceeds by making pairwise comparisons (either by
the decision-makers (DM) themselves or facilitated by analysts interpreting
152
Appendix B: SIMDEC
Appendix B: SIMDEC
153
Fig. B.1 HH-fixed link location at Helsingr (Elsinore)-Helsingborg (from Google Maps)
Table B.1 The four alternatives incl. cost in bn DKK for the HH-fixed link
HH-fixed link (alternatives)
Description (type of construction)
Alternative
Alternative
Alternative
Alternative
A1
A2
A3
A4
Total cost
(bn DKK)
7.7
5.5
11.5
6.0
The four alternatives are listed in Table B.1 with indication of type of
construction and total cost (1 US$ equals around 5 DKK) (Ibid.)
In general the pure person transport alternatives seem to perform best in a
conventional CBA whereas alternatives with rail improvement gain in the wider
SIMDEC analysis. Due to the fixed links high influence on nearby towns, the
impact on the ecology conditions in the sound (tunnels better than bridge solutions
with regard to flow of water), on regional development (trade, work and educationrelated benefits) and the improvement in network (locally, nationally and in an
EU-perspective) the wider set of decision criteria shown below has been adopted.
Criterion 1: Robustness of feasibility (FRA-performance)
Criterion 2: Impact on towns
154
Appendix B: SIMDEC
A1:
A2:
A3:
A4:
High
Middle
Low
1.34
0.43
2.40
3.01
1.15
0.39
2.17
2.63
1.01
0.34
1.94
2.41
Appendix B: SIMDEC
155
A1:
A2:
A3:
A4:
1.34;
0.43;
2.40;
3.01;
37%
0%
86%
100%
Middle
Low
1.15;
0.39;
2.17;
2,63;
1.01;
0.34;
1,94;
2.41;
22%
0%
83%
97%
11%
0%
78%
96%
A1
A2
A3
A4
A1
A2
A3
A4
0
Strong (-6)
Definite (+4)
Strong (+6)
Strong (+6)
0
Strong (+6)
Very strong (+8)
Definite (-4)
Strong(-6)
0
Weak (2)
Strong (-6)
Very strong(-8)
Weak (-2)
0
Note relating to j compared to k: indifference 0, weak +2, definite +4, strong +6 and very strong +8.
Observe that a reversal of j to k is indicated by -. Elements in the diagonal are all necessarily 0.
expected economic growth associated with the actual scenario will affect the
obtained certainty values, see Table B.3.
Where the conventional cost-benefit rate gives a deterministic point estimate of
the feasibility, the RSF-based certainty values give a probability-based interval
estimate of how the two most important uncertainty factors could affect such a
point estimate. Specifically, construction costs and time savings are simulated
using historical reference class knowledge made operational by using the Erlang
and Beta Pert distributions respectively with the latter embedded in a scenario
context (Ibid.). In the simulation the uncertainty due to the estimation of
construction costs is considered generally, i.e. across the scenarios and not related
to a specific scenario.
In SIMDEC the certainty graphs and certainty values are used as the basis of the
final REMBRANDT procedure with regard to criterion 1 about robustness of
feasibility. The four alternatives are compared two by two resulting in altogether
(4 9 3)/2 = 6 pairwise comparisons as shown in Table B.4.
For the remaining four criteria, information has been gathered to serve as
sufficient background for the criteria rating, which leads to additional 4 9 6
comparisons. As an example the pairwise comparisons for criterion 5 about impact
on transport network and accessibility are indicated in Table B.5.
The rating values for all five criteria are shown in Table B.6 together with the
criteria weights, which have been determined by the ROD technique. In case
pairwise comparison had also been applied for the five criteria, this would have
demanded another (5 9 4)/2 = 10 comparisons.
156
Appendix B: SIMDEC
Table B.5 REMBRANDT rating of criterion 5: impact on transport network and accessibility
Pairwise comparison of alternatives under criterion 5: network and accessibility
Scale value
(j, k)
A1
A2
A3
A4
A1
A2
A3
A4
0
Strong (-6)
Definite (+4)
Weak (+2)
Strong (+6)
0
Very strong (+8)
Strong (+6)
Definite(-4)
Very strong (-8)
0
Definite (-4)
Weak(-2)
Strong (-6)
Definite (+4)
0
A1 rates
A2 rates
A3 rates
A4 rates
Weights
Robustness of feasibility
Impact on towns
Impact on ecology
Impact on regional economics
Impact on network and accessibility
0.03
0.20
0.47
0.05
0.05
0.00
0.79
0.47
0.00
0.00
0.19
0.00
0.03
0.76
0.84
0.78
0.01
0.03
0.19
0.11
0.35
0.13
0.06
0.19
0.27
Table B.7 The four alternatives with total scores and rank order indicated
HH-fixed link Description (type of construction)
Cost
(alternatives)
(bn DKK)
Score
Rank
A1
A2
A3
A4
0.09
0.13
0.44
0.34
4
3
1
2
7.7
5.5
11.5
6.0
Appendix B: SIMDEC
157
B.4 Discussion
Conventional decision support for deciding upon the four HH-alternatives would
consist of a calculation of benefit-cost rates (BCRs) supplemented by various
information not accounted for in the BCRs. SIMDEC offers an approach where
simulation and MCA are applied to deal with the complex decision problem. The
BCR information contained in Table B.2 indicates that three (A1, A3 and A4) out
of the four alternatives are socio-economically sound with BCR values also in the
low growth scenario being around or above 1. Conducting reference scenario
forecasting (RSF), however, indicates, by inspecting the produced certainty graphs
(CGs) and certainty values (CVs) that really only two alternatives (A3 and A4) are
sound when including RA in the assessment (CV C 7580%). By accounting for
estimation bias in the cost estimates and prognosis bias in the forecasting of traffic
and exploring this by simulation embedded in scenarios, the alternative A1 is
shown to have a feasibility that cannot be considered robust.
With alternatives A3 and A4 remaining as candidates for implementation, these
are explored in a wider context where strategic, non-monetary issues are
introduced together with the criterion about robustness of feasibility based on
the described risk analysis. With CGs and CVs of the alternatives as input to this
criterion a REMBRANDT multi-criteria analysis is carried out which comprises
also the criteria about the impacts on towns, on ecology, on regional economics
and on transport network and accessibility. The result is that even with the highest
criterion weight on robustness of feasibility the order of importance with regard to
A4 and A3 is now reversed as alternative 3 now becomes the most attractive
alternative.
For the lower ranking alternatives A1 and A2 it should be observed that A2 is
now better than A1. The wider assessment based on the multi-criteria analysis has
thus revealed and indicated some qualities contained in A2, which were not
captured by the BCR values.
158
Appendix B: SIMDEC
priority-ordering of the alternatives. One important issue that can be treated is how
different stakeholder strategiesdefined by the chosen set of decision criteria and
the stated value inputs to the pairwise comparisonswill affect the result (Leleur
2008; Jeppesen 2010).
The SIMDEC approach has so far been tested on transport planning problems
but the perspective is to explore its potential also for complex decision problems
outside the transport sector. It is expected that alternatives for construction projects
in general can be examined by SIMDEC in a way that satisfies both theoretical
validity and practical userfriendliness.
Appendix B: SIMDEC
159
AHP
REMBRANDT
Indifference
2, 4, 6, 8
1, 3, 5, 7
Weak
Definite
Strong
Very strong
Compromise
There are as mentioned three main criticisms of the AHP method which the
REMBRANDT method tries to correct. The first is related to the scale in AHP
where 1 represents two objects being equal in value, 3 means that the first object is
slightly better than the second object, 5 indicates clear preference in this respect, 7
a strong preference and 9 a very strong preference. Based on a number of examples
and reflections, Lootsma (1999) has adjusted the numerical scale for
REMBRANDT, so it is more convenient for subsequent calculations.
The second point that REMBRANDT tries to improve is the calculation of
scores. AHP uses a method which has the disadvantage that if a new alternative is
added later in the process, it may reverse the existing ranking of alternatives
(known as rank reversal of alternatives). REMBRANDT uses logarithmic
regression or geometric mean, whereby the potential problem of rank reversal is
overcome. For a more detailed technical analysis refer to Olson et al. (1995).
The third and last point which the REMBRANDT method tries to improve
compared with AHP is the way the individual scores are aggregated. The AHP
uses a method based on calculation of eigenvectors leading to scores and
arithmetic mean aggregation by summation of the scores multiplied by the criteria
weights, while REMBRANDT calculates the value of an alternative by using the
geometric mean scores and multiplying these scores after they have been uplifted
with the criteria weights.
To illustrate the principles of REMBRANDT a small calculation example is
described below (Ibid.).
There is a decision problem involving three alternatives (A, B and C) and four
criteria (W, X, Y and Z). The criteria weights are already set to:
0:493; 0:246; 0:174; 0:087
Scores for each alternative under each criterion is calculated using the following
transformation: eln(2) d(jk). It is noted that when REMBRANDT is used to
determine criteria weights the transformation eln(H2) d(jk) is used (Lootsma 1999;
160
Appendix B: SIMDEC
Barfod and Leleur 2011). The pairwise comparisons of the three alternatives under
each of the four criteria are shown below:
Pairwise comparison:
Transformation:
Criterion W:
A B
A 0
4
B -4 0
C -6 -4
C
6
4
0
Criterion X:
A B
A 0
-2
B 2
0
C -1 -4
C
1
4
0
A
B
C
Criterion Y:
A
B
A 0
0
B 0
0
C 4
3
C
-4
-3
0
Criterion Z:
A
B
A 0
1
B -1 0
C 1
2
C
-1
-2
0
B
16
1
0.0625
C
64
16
1
10.08
1
0.0992
A
1
4
0.5
B
0.25
1
0.0625
C
2
16
1
0.7937
4.0
0.3150
A
B
C
A
1
1
16
B
1
1
8
C
0.0625
0.125
1
A
B
C
A
1
0.5
2
B
2
1
4
C
0.5
0.25
1
A
B
C
A
1
0.0625
0.015625
Geometric mean:
0.3969
0.5
5.0397
1
0.5
2
Afterwards the total score for each alternative A, B and C is found by using the
determined values above and the criteria weights of the four criteria W, X, Y and Z
as indicated below; the obtained total scores are as a matter of convention (like in
AHP) transformed into a normalised set of numbers.
A:
B:
C:
10.080.493
10.493
0.09920.493
* 0.79370.246
* 40.246
* 0.3150.246
* 0.39690.174
* 0.50.174
* 5.03970.174
* 1 0.087
* 0.50.087
* 20.087
=
=
=
2.513
1.174
0.339
0.624
0.292
0.084
Appendix B: SIMDEC
161
The use of MCS for transport evaluation is described in (Salling 2008) with a
special focus on feasibility risk assessment (FRA) concerned with the examination
of the certainty that a given transport investment project is feasible from a socioeconomic viewpoint. This can be illustrated as follows.
The CBA result is typically presented by one or more of the following index
values: the net present value (NPV), the internal rate of return (IRR), the benefitcost rate (BCR) or a combination of these. Setting focus on the BCR value the
MCS is used to determine the robustness of feasibility when a CBA-based point
estimate is transformed into an interval result provided by a certainty graph for the
investment showing non-feasible and feasible outcomes by use of probability
estimates of the BCR values. Specifically, this graph is made up of the probability
estimates of achieving at least the BCR indicated as argument; thus
CG(x) = Prob(BCR C x). An example is shown in Fig. B.3 indicating a 90%
confidence interval between BCR-rate values 0.80 and 2.44 and a certainty value
(CV) equal to 83%, which means that the BCR-rate value has a 83% probability of
being equal to or higher than the cut-off value equal to 1.
In recent research of feasibility of transport investment projects it has been
verified that especially construction costs and traffic demand forecasts are
important (Priemus et al. 2008; Salling and Leleur 2009; 2010).
In an ongoing research project Uncertainty in Transport Project Evaluation,
UNITE (20092012), funded by the Danish Strategic Research Council a decision
support system (DSS) has been developed containing both a deterministic CBA
module, a deterministic MCA module (AHP and REMBRANDT) and a stochastic
MCS module. Thereby the DSS software can support both the COSIMA
162
Appendix B: SIMDEC
Appendix B: SIMDEC
163
Belton V, Stewart TJ (2002) Multiple criteria decison analysisan integrated approach. Kluwer
Academic Publishers, Dordrecht
Bhushan N, Rai K (2004) Strategic decision making: applying the analytic hierarchy process,
springer series in decision engineering
DMG (2010) Decision modelling group compendium decision support: theory and practice,
DMG, DTU Transport. Technical University of Denmark, Denmark
DTU (2007) Dokumentationsmodel for TGB Vurderingsmodel (Trafikplan for Grnland:
Beslutningsredskab), Center for Trafik og Transport (CTT). Danmarks Tekniske Universitet
ECMTEuropean Conference of Ministers of Transport, Group of Experts (1981) Exchange of
information criteria applied to transport infrastructure projects, ECMT Publication
Expert Choice Inc (2004) Expert Choice, Arlington
Goodwin P, Wright G (2010) Decision analysis for management judgement, 4th edn. Wiley, New
York
Hiselius LW, Barfod MB, Leleur S, Jeppesen SL, Jensen AV, Hjalte K (2010) Helhetsorienterad
utvrdering av kollektivtrafiktgrder, Bulletin 246, Institutionen fr Teknik och Samhlle,
LTH, Lunds Universitet
Hwang CL, Yoon K (1995) Multi attribute decision makingan introduction, Sage University
Series No. 07-104, Sage Publications, London
Jensen AV, Barfod MB, Leleur S (2007) Strategic planning and decision analysis: presentation of
the COSIMA software system. Paper presented at the INFORMS 2007 Congress, Puerto Rico,
July 2007
Jeppesen, SL (2010) Sustainable transport planning. Ph.D. dissertation, Department of Transport,
Technical University of Denmark
Kronbak J (1998) Trafikplanlgning og GIS-baserede konsekvensberegninger. Ph.D. Dissertation, Institut for Planlgning. Danmarks Tekniske Universitet
Larsen LA, Skougaard BZ (2010) Vurdering af alternativer for en fast forbindelse HelsingrHelsingborg. M.Sc. thesis, DTU Transport, Danmarks Tekniske Universitet
Leleur S (2000) Road infrastructure planninga decision-oriented approach, 2nd edn.
Polyteknisk Forlag, Lyngby
Leleur S (2008) At navigere mod fremtiden: Systemisk planlgning som ide og metode.
Polyteknisk Forlag, Lyngby
Leleur S, Holvad T (2004) New appraisal methodology for large European transport investments.
In: Proceedings of the European transport conference 2004. Association for European
Transport (AET), Strasbourg
Leleur S, Jensen AV, Salling KB (2004a) COSIMAsoftware manual, Version June 2004,
Centre for Traffic and Transport (CTT). Technical University of Denmark
Leleur S, Jensen AV, Salling KB (2004b) Modelling decision support and uncertainty for large
transport infrastructure projects: the CLG-DSS model of the resund fixed link. In: Antunes
CH, Dias LC (eds) Proceedings of the 15th mini-EURO conference in managing uncertainty
in decision support modelsMUDSM 2004, INESC-Coimbra, Coimbra, 2004
Leleur S, Petersen NB, Barfod MB (2007) The COSIMA approach to transport decision making:
combining cost-benefit and multi-criteria analysis for comprehensive project appraisal. In:
Kim K-S (ed) Proceedings of the Korean Development Institute and World Bank conference,
Seoul, May 2007. Improving public investment management for large-scale government
projects: focusing on the feasibility studies. KDI, pp 100122
Leleur S, Larsen AL, Skougaard BZ (2010) Strategic transport decision making: the SIMDEC
approach based on risk simulation and multi-criteria analysis. In: EUROSIS Proceedings of
the Asian simulation technology conference ASTEC 2010 conference, Shanghai, March 2010
Lootsma FA (1999) Multi-criteria decision analysis via ratio and difference judgement. Kluwer
Academic, Dordrecht
Olson DL (1996) Decision aids for selection problems, springer series in operations research
Olson DL, Fliedner G, Currie K (1995) Theory and methodology: comparison of the
REMBRANDT system with analytic hierarchy process. Eur J Oper Res 82:522539 (Elsevier)
164
Appendix B: SIMDEC
Palisade (2007) @RISK version 5.0Advanced risk analysis for spreadsheets. Palisade
Corporation, New York
Priemus H, Flyvbjerg B, van Wee B (2008) Decision making on mega-projects: cost-benefit
analysis, planning and innovation. Edward Elgar Publishing, Cheltenham
Roberts RJ, Goodwin P (2002) Weight approximations in multi-attribute decision models.
J Multi-Criteria Decis Anal 11:291303 (Wiley)
Saaty TL (2001) Decision making for leaders: the analytic hierarchy process for decisions in a
complex world. RWS Publications, Pittsburgh
Salling KB (2008) Assessment of transport projects: risk analysis and decision support. Ph.D.
Dissertation, Department of Transport, Technical University of Denmark
Salling KB, Leleur S (2009) Modelling of transport project uncertainties: risk assessment and
scenario analysis. In: Proceedings of the modelling and applied simulation09 conference,
Puerto de la Cruz, Tenerife, Sept 2009
Salling KB, Leleur S (2010) Reference scenario forecasting: a new approach to transport project
assessment. Paper presented at the 12th world conference on transport research in Lisbon, July
2010
Steffensen M, Testmann L (2000) Vurdering af omfartsveje belyst gennem alternative projekter
for Hng. M.Sc. thesis, Institut for Planlgning, Danmarks Tekniske Universitet
Trafikministeriet (2003) Manual for samfundskonomisk analyse, Copenhagen, juni 2003
Tsamboulas D, Beuthe M, Grant-Muller S, Leleur S, Nellthorp J, Panou K, Pearman A, Rehfeld C
(1998) Innovations in decision analysis, deliverable D10, EUNET research project, EU 4th
Framework Programme
Vaidya SO, Kumar S (2006) Analytic hierarchy process: an overview of applications. Eur J Oper
Res 169:129 (Elsevier)
Vose D (2002) Risk analysis: a quantitative guide, 2nd edn. Wiley, New York
von Winterfeldt D, Edwards W (1986) Decision analysis and behavioural research. Cambridge
University Press, Cambridge
Index
@ RISK, 161
A
Ackoff, 43
AHP, 65
Airport in Greenland case, 100
Analyst, 96
Analytic hierarchy process (AHP), 52, 62,
80, 108
Appraisal study, 125, 140
Arationality, 38
Aristotle, 23
Artificial intelligence, 37
Ashby, 19
Assessment, 55
Autopoeisis, 15
B
Benefit-cost rate (BCR), 64
Bhushan, 109
Black Swan theory, 113
Bohr, 2930
Boundary setting, 7, 57
Brainstorming (BS), 52, 72, 90, 103
Buchanan, 6
Business innovation, 117
Business problems, 108
Butterfly Effect, 8
C
Capra, 2930
Causality, 23
165
166
C (cont.)
COSIMA software, 136, 144
Cost-benefit analysis (CBA), 52, 6263, 92
Criteria weights, 84
Critical systems heuristics (CSH), 44, 52,
5758, 76, 90, 98
Critical systems thinking (CST), 44
Customised decision model, 136137
D
Danish bypasses case, 97
Danish Strategic Research Council, vii
Decisions, 20
Decision analysis, 7677, 140
Decision analyst, 77
Decision analysts, 13, 22
Decision awareness, 115
Decision conference, 73, 76, 95, 135
Decision conference participants, 92
Decision criteria, 124
Decision engineering, 109
Decision ownership, 115
Decision space, 57, 73, 115
Decision support, 2, 66, 77, 106, 124, 150
Decision tree, 80, 142
Detached understanding, 96
Detail complexity, 5
Deterministic point estimate, 155
Direct pricing, 64
Dreyfus and Dreyfus, 31, 3538, 68
Dynamic complexity, 8
E
EcoMobility (20102012), vii
EcoMobility assesment model, 162
Economics, 64
Economic assessment, 63
Economic consequences, 64
Eigenvector problem, 143
Ellis, 117
Emancipatory mode of enquiry, 46
Emancipatory paradigm, 46
Emergency planning, 108
Energy planning, 106
Epistemic lenses, 30
Epistemology, 27
EU Regional Development Fund, 162
European strategic transport
research, 106
EU transport studies, 124
Expert Choice, 143
Index
F
Facilitator, 90, 96
Financial analysis (FA), 63
Five-stage learning model, 37, 96
Fixed link between Elsinore
and Helsingborg case, 107
Flood, 7
Flyvbjerg, 151
Foucault, 44
Functionalist mode of enquiry, 45
Functionalist paradigm, 45
Futures workshop (FW), 52, 57, 90
G
Generic decision making
problem, 108
Geographic information
system (GIS), 76
Gilboa, 68
Globalised world, 1
Glocalisation, 1
Goodwin, 141
Governance problems, 108
Group learning, 76
Group processes, 68, 7677, 95
H
Habermas, 10, 44
Hard methods, 52, 62, 107, 116
Hardin, 9
Heisenberg, 2930
High speed rail case, 102
Holistic, 2, 29, 39
Horizon of possibilities, 112
I
Information technology, 76
Intelligent transport systems
(ITS) case, 105
Interactions, 17
Interpretation, 95
Interpretive mode
of enquiry, 45
Interpretive paradigm, 45
Investment criteria, 64
Involved understanding, 96
J
Jackson, 4348
Index
K
Kahneman, 151
Khisty, 12
L
Law of Requisite Variety, 19
Learning, 95
Leleur, 47, 52
Light rail service case, 103
Linstone, 48
Litmus test, 97
Log book, 83, 106, 147, 156
Long-term consequences, 90
Lootsma, 158159
Lorenz, 8
Luhmann, 15, 1720, 31, 68, 115
Lyotard, 44
M
Management, 20, 23
Management thinking, 67
Mandelbrot, 5
Market pricing, 64
Martin, 117118
Maturana, 16
MCA information, 106
McCarthy, 10
Means-ends configuration, 12
Midgley, 115
Mind mapping (MM), 52, 57, 104105
Mitroff, 48
Modes of enquiry (MOEs), 27, 4547, 95
Mohammadi, 12
Monte Carlo simulation, 66, 92
Morin, 29, 31, 115
Multi-attribute utility theory, 65
Multi-criteria analysis (MCA), 64, 92
Multi-methodology approach, 109
Multiple perspectives, 48
N
Narayanan, 118
Net present value (NPV), 64
Newtonian physics, 28
Non-economic
consequences, 64
O
Olson, 141, 158159
Open-ended change, 21, 27
167
Operations research, 39, 64
Optimisation, 39, 64
Organisation, 20
Organisations, 17
Organisational decision making, 63
Organised complexity, 10
Ownership of the problem, 97
P
Pairwise comparison, 65, 84, 103, 157
Paradigms, 27
Perception of learning, 35
PerronFrobenius theory, 143
Planners, 13, 22
Planning, 1819, 22, 49, 56
Planning team, 48, 50
Poincar, 8
Postmodern mode of enquiry, 46
Postmodern paradigm, 46
Preference analysis (PA), 52, 62, 65
Preference complexity, 10, 112
Preference information, 81
Prigogine, 23
Proactive effort, 22
Probability-based interval, 155
Problem-solving, 49
Project appraisal, 123
Project feasibility, 124
Promoting biking in
Denmark case, 104
Q
Quantitative risk analysis, 137
Qvortrup, 1
R
Rai, 109
Rank order distribution (ROD)
weights, 84, 126
Rationality, 38
Recasting of systemic perceptions, 48
Reframing/back-talk schema, 32
REMBRANDT, 65, 143, 151, 158159
Risk, 65
Risks, 62
Risk analysis, 66, 92
Risk analysis based on Monte Carlo
simulation (RA), 52, 62
Robustness, 116, 153, 157, 161
Rosenhead, 23
Rumsfeld, 112
168
S
Saaty, 142
Scanning, 55
Scenarios, 154
Scenario analysis (SA), 52, 62
Schn, 3132
Scoping, 5556
Search-learn-debate process, 39
Self-organisation phenomena, 24
Self-organising multi-causality, 28
Senge, 7, 9, 11
Shannon, 10
SIMDEC, 106, 151, 156
SIMDEC approach, 149
Simon, 10, 36
Simple multi-attribute ranking technique
(SMART), 52, 62, 141
Simplicity paradigm, 28
SMART technique, 141
Social return of investment (SROI), 117
Social systems, 1617, 19
Societal complexity, 1
Socio-technical system, 5, 7, 112
Soft methods, 52, 57, 72, 90, 116
Soft systems methodology (SSM), 52, 57, 59,
76, 90, 101
SP framework, 2, 69, 109, 118
Stacey, 8, 2123, 68
Stakeholder analysis (STA), 52, 57, 65, 72
Stakeholder viewpoint, 90
Step-by-step approach, 49
Strategic cognition (SC), 118
Strategic decision making, 56
Strategic decisions, 2
Strengths, weaknesses, opportunities and
threats (SWOT), 52, 57
Subnuclear physics, 28
Suboptimisation, 57
Subworld, 38, 76, 95, 115
Sustainability, 117
SWOT analysis, 90
Swing weight (SW) method, 144
Sympoietic, 24
System, 16, 47, 49
System complexity, 19
System demarcation, 7
System environment, 17
System/environment, 15
Systematic approach, 2, 4950
Systematic assessment, 99
Systematic method-elements, 2
Systematic planning, 2, 24, 72
Systematic thinking, 30
Index
Systemic perception, 48
Systemic planning (SP), 95
Systemic scanning, 99
Systemic thinking, 30
Systemic toolbox, 50, 52, 76, 9091
Systems analysis, 49
Systems science, 28, 43, 49
Systems techniques, 39
Systems theory, 16
Systems thinking, 16, 44
T
Taleb, 113114, 117
Theory of communication, 10
Thyssen, 17, 20
Total rate of return (TRR), 87
Total systems intervention (TSI), 44
Trade-off analysis, 85
Tragedy of the commons, 9
Transformative teleology, 23
TRANS-IT Consult, 71, 92
Transport decision making, 124, 137
Transport infrastructure planning, 124
Transport modelling, 154
Transport planning, 106, 116
Tversky, 151
Types of change, 21
U
Ulrich, 44, 58
Uncertainty, 1, 29, 65, 117
Uncertainty principle, 30
UNITE (20092012), vii
Unknown unknowns, 2, 112
V
Value function (VF), 79
Varela, 16
Vessel traffic service case, 101
Von Koch, 5
W
Weight sets, 142
Weights, 64
Wider performance, 57, 75, 88, 91, 97
World complexity, 39
Wright, 141