1994 INTERNATIONAL SYSTEM DYNAMICS CONFERENCE
Innovations in Conceptual Modelling
Marcel M H Quanjel Ivo Wenzler
Policy Advice Group (BAG) Policy Advice Group (BAG)
Institute for Applied Social Sciences (ITS) Institute for Applied Social Sciences (ITS)
P.O. box 9048, 6500 KJ Nijmegen P.O. box 9048, 6500 KJ Nijmegen
the Netherlands the Netherlands
Tel: +-3 1-80653500, Fax: +-31-80653599 Tel: + 31-80653598, Fax: +-31-80653599
E-mail u800045 @vm.uci.kun.nl E-mail u800041 @vm.uci.kun.nl
Abstract
In this paper we describe the process of building a conceptual model of a guided missile base of the
Royal Netherlands Air Force, a complex organisation in the middle of transformation. Based on this
conceptual model we developed a policy exercise that is used by the Air Force to explore their future
organisation, focusing especially on the communication structure.
We will describe the steps made in developing the model and explain the choices that lead to some
methodological innovations in communicating complexity through a conceptual model. We will also
describe this model, which consists of 28 actors or actor groups, and between them more than 350
relationships of nine different types, including five specified types of communication.
Furthermore we will look shortly into the possibilities of transforming our qualitative model of actors
and relations between actors into a system dynamic model, thus broadening the scope of system
dynamics by giving an input from the system analytical approach we use in developing policy
exercises.
Industry, page 85
1994 INTERNATIONAL SYSTEM DYNAMICS CONFERENCE
Innovations in Conceptual Modelling
The problem
The Royal Netherlands Air Force is a rapidly changing organisation. Due to the recent changes in the
European military theatre the decision has been made to withdraw the two Dutch guided missile bases
from Germany. These two will be combined into the new Guided Missile Base (GGW De Peel) in the
Netherlands. The Air Force is also changing its military strategy from a 24 hour alert into the new
NATO strategy of few Reaction Forces that can be in action anywhere in the world on a very short
notice and Main Defence Forces for the times of war.
The withdrawal and joining of the two bases has prompted a number of other processes of change,
mostly initiated by the Working group for the Operational and Logistical Philosophy (WOLF).
According to the WOLF concept the base will change from a matrix organisation to an organisation
based on a system of integral management. In the old management system the base had three
hierarchical levels, in De Peel there will be two. Squadrons will be treated as business units, they will
be responsible for achieving their own production targets, and they will have to negotiate with each
other over potentially scarce resources. On top of this all production p will be led by a
system of process management and integral quality assurance.
The large number of processes of change have some significant consequences. For the people
involved it leads to:
* — loss of oversight,
* — lack of knowledge, and
* feeling of uncertainty.
Therefore the devel of the new isation is ied by a lot of discussion in which
stakeholders participate from:
¢ their specific backgrounds,
« their differing levels of knowledge, and
+ their specific interests.
To reduce inty about the ication structure and to come to a more fruitful development
of the new organisation, the command of GGW De Peel ‘and the Department of Behavioural Sciences
of the Air Force Staff have chosen for an i and p devel of the
communication structure in the new base. This process is facilitated by the Policy Advice Group
(BAG) of the Institute for Applied Social Sciences (ITS).
The project
The main objective of the project was to develop an approach that leads to a clear and functional
communication structure for GGW De Peel. The process consists of three different phases. Phase 1
started in September 1993 and ended in January 1994. Phase 2 started in January and ended in May.
Phase 3 will end in November 1994.
Industry, page 86
1994 INTERNATIONAL SYSTEM DYNAMICS CONFERENCE
Creating a shared vision (phase 1)
Objective + In this initial phase our objective was to create clarity in the discussion about the new
organisation and to develop a shared vision among stakeholders on the way the new base is supposed
to function.
Process + We therefore developed a structured process where key actors of the base and a number of
officers from the Air Force Staff jointly defined the tasks and responsibilities of the main actors in the
new base. They also defined the relations between these actors and they identified a number of
potential probl in the they redesigned. They also developed several solution scenarios
that could respond appropriately to these problems.
Result + The findings were reported in the 'Systeemanalyse GGW De Peel' report, and the conceptual
model that is part of this report.
Testing the new communication structure (phase 2)
Objective + In this second phase our objective was to test the structure that was defined during the
previous phase of the project.
Process * Based on the conceptual model that resulted from the previous phase we developed a policy
exercise (Brewer 1986, Wenzler 1993) in which the communication aoucnire of the new base could
be tested. In May of this year 37 key persons of the new base i d in a simulated
the new communication structure they defined themselves. They were also able to experiment in a
safe environment with their own responses to the potential problems in the new structure.
Result + Based on the experiences of this policy exercise the command of the new base was able to
approve those parts of the new structure that proved to be adequate, and to change the communication
structure where necessary.
Training the squadron management (phase 3)
Objective * Until now only the top management of the new base was involved in the development
process. There is also a need to the of the decisions made in the two
previous phases to the middle levels in the i
Process + For the above objective to be achieved we are ly developing a gaming/simulati
exercise focusing on the communication structure of the new base, that will involve the middle
management levels in the process.
Result * On the one hand the result is an inventory of problems that can be expected on the middle
management level, and on the other hand it is an opp: ity for middle to experiment
with p ial soluti to those probl Another result is the transfer of the communication
Industry, page 87
1994 INTERNATIONAL SYSTEM DYNAMICS CONFERENCE
structure that was developed in phase | and 2 to the lower levels in the organisation. In this respect
phase 3 is a logical follow up for the two previous phases.
In this paper we will elaborate on phase | of the WOLF project: the conceptual modelling and its
result, and the visualisation of a shared mental model of the communication structure of GGW De
Peel.
Conceptual modelling
Knowledge about the new structure
When phase 1 of our project started we had no specific knowledge about the army. So our first task
was to learn what the Air Force did, what their language was, and what types of processes they have.
We did this by making a thematic inventory of policy documents (step 1) that our client, the Air Force,
designated as central for our project. This was done by grouping pieces of text per actor and per
process on thematic cards, thus creating a database of everything that was stated in these documents
about all the central elements of the WOLF concept and its relation to the new guided missile base.
We grouped all thematic cards on a large matrix on the wall, making distinction between different
levels in the Air Force (NATO. Air Force, GGW De Peel, Squadrons). In doing so we were able to
learn enough about WOLF. the Air Force and the base itself to be appropriate discussion partners on
the issue. We were now able to ask the right questions during the interviews with the key actors (step
2) of the new base. The central theme of the interviews, and the subsequent discussion, was the
communication structure of the base: how can the C der and his staff i with 9
independently operating squadrons without losing control and without violating the WOLF-concept?
The following task was to translate our knowledge about the communication structure of the base into
a conceptual model, described in terms of actors and relations between those actors.
Conceptual model
Together with our client we first decided on the level of abstraction (step 3) we would use in the
conceptual model of the base. It was decided that squadrons and staff departments in the base would
be regarded as unitary actors. In total we distinguished 9 squadrons and 19 staff actors.
As in most organisations, the Air Force uses hierarchical structures to visualise its own organisation:
the i However. in visual an isational problem it is much more useful to think in
terms of a functional structure (step 4). We started developing the ional around the
primary process of the base, the production of air defence. Air defence is produced by four squadrons
(Triads). Each TRIAD go through four different phases (status of alertness) in a cycle of two years: 1)
ready to go, 2) maintenance, 3) nearly ready to go, and 4) training. By the nature of their status they
have a special orientation towards other sq and staff d For le: if a Triad has
a maintenance status, they have scheduled most of their equi for maii within this period.
Industry, page 8
1994 INTERNATIONAL SYSTEM DYNAMICS CONFERENCE
Of course their relationship with the logistical dron is domi while their relationship with the
staff d that itors its ional perfc is of minor importance.
This was an important aspect of the new military strategy that guided us in giving the conceptual
model its shape. We grouped all squadrons and staff departments around the four TRIAD squadrons,
using the functional orientation as a guide. After designing this initial structure, we proceeded as we
were used to in our previous projects: we attached a 2*2 m sheet of paper to the wall, drew the actors
(step 5) conform their functional orientation and started filling the actors in the model with literally all
the information we had about them.
Of course this information is often formulated in terms of a relations with other actors (step 6) . If so,
we connected actors with each other. Very soon we were no longer able to read our own model,
because it contained more then 175 lines between the actors. Because it was not possible to follow the
line itself, and to keep grip on the inf ion, we wrote at the beginning of each line its destination.
We no longer looked at the line itself, but the text at the beginning of the line told us at which actor to
look. We then realised that this is an important concept to deal with large quantities of information.
We developed this logic further: if you know the destination of a relationship line, you don’t need the
line itself, but you could use abstractions of this line. These ab i we call them higt
could be used to represent something else, for instance the communication structure of the base. More
over: if we give information about the destination of a relationship line, we could also attach
information about the content of that relationship. We regarded this as the first innovation in our
process of conceptual modelling.
According to this concept we designed a pattern of ‘relation highways' in the functional structure of
the new base. From each actor relations enter this highway. The link to the highway is made of labels
that contain information about the destination of the relation and of the character of the relation (step
7). By giving each actor a logical number, we could make use of the number to identify the actor. By
making use of icons, we were able to distinguish between as many types of relations as we wished,
and moreover: the type of the relation is known at first sight. We regard this our second innovation in
making use of visual language in a conceptual model. In figure (1) we show the basic structure of the
conceptual model, and in figure (2) an example of one of the actors and its links to the relation
highways.
Industry, page 89
1994 INTERNATIONAL SYSTEM DYNAMICS CONFERENCE
Figure | * The basic structure of the conceptual model of GGW De Peel
Industry, page 90
1994 INTERNATIONAL SYSTEM DYNAMICS CONFERENCE
Ca)
(RB)
> \ ge -\ a5) Nae ‘ge
LE CF LE @
SP LE LOW D LY
® 2 ib)
24 ¢ Operational
Squadron }
> DD O © 6)
ef e 208.
~ a fon)
23 23 27 26)
Figure 2 * Example of an actor
(different types of relationships are linked to a number of actors)
To represent the nature of a relations we used 9 different icons, 5 of them representing specific types
‘ched
ion. We disti
an order, an advice, information, mutual adjustment or
negotiation, a request, people, money, a product and monitoring. For each of the 9 types of relations
we attached the destination by using the number of that actor. This concept made it possible to
Industry, page 91
1994 INTERNATIONAL SYSTEM DYNAMICS CONFERENCE
on a very effective and easy id way 28 actors and over 350 relations between
them, while each relation can hold a maximum of 9 specific types of that relation. Although this was
not our main objective, this way of representing a model proved to be also very elegant and visually
attractive.
We also used 5 different colours in the conceptual model, each of them representing a different
functional cluster in the organisation. Use of colour coding (our third innovation ) made it easy to see
where the destination of a relationship is. In a first glance you see the functional cluster by its colour,
and you recognise the specific actor in that cluster by its number. This way of representing the
relations with other actors in an organisation has an important benefit: it gives an immediate insight in
the nature of the function of an actor. Some actors have contacts only in their own sector, which is
indicated by the fact that almost all their relationships have their own colour. Other actors are real
spiders in a web: their contacts are similar to a rainbow.
Developing a shared vision
All the information we gathered in our research would have had no extra value for the organisation if
it would not be internalised by the stakeholders involved. Therefore it was of the utmost importance to
rebuild this whole model with the actors that are rep din this ptual model. We designed a
structured two day workshop where in a first step the central tasks and responsibilities of every actor
in the model were red d and di: d. Thus all the partici in the workshop agreed on who
was doing what in their new organisation. In a second step they split up in several functionally
oriented groups and we asked them first to redefine all the relations between actors and then give a
detailed description of each relation. These relations were also discussed in plenary sessions. All the
information we gave them in these two days was thus reconsidered, discussed and agreed upon. Our
model had become much richer, much more detailed and above all, it had become their model!
Beyond our own tradition
The conceptual modelling process, as described above, serves for our group primarily two purposes.
The first purpose of the process is to provide us with a tool to help our clients in developing a shared
vision of their reality. The second one is to provide us with a detailed conceptual model of our client's
reality which is then used for developing a ‘simulated reality'. Most of the times this 'simulated reality’
we develop is a policy exercise or a gaming/simulation exercise.
In the recent tradition of system dynamics, the conceptual modelling process also basically serves the
above ioned purp Developing causal di: ( 1 modelling) with the client has
been proven effective in helping them develop a shared vision of their reality (Vennix 1993). Causal
diagrams are also used as a basis for developing ‘simulated realities’, which in the language of system
dynamics means dynamic simulation models and micro-worlds.
Although these two approaches are more than significantly similar and complementary (Figure 3),
there has been very little attempts to make a connection and learn from each other. Our belief is that
our (qualitative) conceptual modelling process, which focus on actors and their relations, could be
Industry. page 92
1994 INTERNATIONAL SYSTEM DYNAMICS CONFERENCE
very useful not only in developing the itative) causal di of complex probl but also in
the process of d ping itati Iti-di ional dynamic sii ion models. At the same
time we also believe that we can benefit a great deal from the modelling tradition of system dynamics,
especially for developing complex policy exercises which need to have sophisticated and dynamic
decision-support systems as part of their design.
System dynamics Policy exercises
f
1
Conceptual | Causal Functional
modelling | diagrams (actor and
i relations ) model
|
J
Dynamic Policy a
Simulated . simulation exercises an
realities | models and gaming/
micro- worlds fous ,
itati qi tative an
(quantitative) uanilisive
Figure 3 * System dynamics and policy exercise as complementary approaches
Bibliography
Brewer, G.D., Methods for policy synthesis: policy exercises, in: Clark & Munn (eds.), Sustainable
development of the bio sphere (pp. 455-473), Cambridge, 1986
Vennix, J.A.M., W.J. Scheper & R. Willems, System Dynamics: What does the client think of it?,
Paper presented at the International System Dynamic Conference 1993 in Mexico
Wenzler , I., Policy Exercises. A New Approach to Policy Development, Nijmegen, 1993
Industry, page 93
