Table of Contents
The A pplication of Influence Diagrams for the Development
of Military Experiments
Mr Duncan Tailby (Presenter)
Defence Science and Technology Organisation, Australia
PO Box 1500, Edinburgh South Australia 5111
Ph: +61-8-8259- 5817 Fax: +61- 8-8259-5624
duncan.tailby@dsto.defence.gov.au
Prof Geoff Coyle
Visiting Professor of Strategic Analysis,
Univesity of Bath, UK
geoff.coyle@btintemet.com
Dr Andrew Gill
Defence Science and Technology Organisation, Australia
PO Box 1500, Edinburgh South Australia 5111
Phr +61-8-8259- 5112 Fax: +61-8-8259-5624
andrew. gill@dsto.defence.gov.au
Abstract
The use of qualitative, diagrammatic, modelling is a growing trend in system dynamics
with a well-developed literature, designed to deal with problems with a multitude of
‘soft’ variables, strong influences from human actors, and profound effects from
political forces. An application area is the design of defence experiments to assess the
effects of contemplated military investments on the ability to meet national goals. The
difficulty lies in the design of experiments to test the system under a range of reasonably
realistic, but sufficiently different, contexts to provide a comprehensive understanding
of the issues. The paper will describe a hypothetical instance of tensions between
nations and develop an influence diagram that inter-relates the political factors with the
military factors. Study of the feedback loops connecting the political and military
factors show how distinctly different experiments can be designed to test the totality of
national crisis management capabilities.
Key words: Influence diagram, defence, design of experiments, strategic planning,
scenarios.
1. Introduction
1.1. Background
Our futre global security environment is characterised by uncatainty. Such om
emvironment presents many challenges for staegic plames oconsidaing enaging
social, political, military and technological trends and the events they may lead to. A
goal of military strategic planning is to understand wher these trends may lead and
prepare for the potential eventalities This paper will eqlor the tecmiqe of
quditative, diagrammatic modeling for andlysing such eventalities to infom nilitay
capability devdopment.
Military strategic planning begins with understanding the key factors and characteristics
of the range of potential fubwe environments and the likely ares of friction Actions
taken by nations to affect key factors can be classified into four domains: diplomatic,
infomation, military, and economic (Austalian Department of Defence 2002, 11).
Action in any one, or a number of the domains is intended to either reduce the tension
within the environment or to cree conditions favourable to a nation The advesaial
nature of some of the national actors produces a situation where the environment can be
characterised by a number of key factors and variables that are in constant transition
within a complex ooartive system (Flood and Jackson 1991, 32; Groom 2002, 4). The
difficulty that aisss in examining sich ewioments for th puposs of nilitay
capability development is that many of the variables ae soft and the influences they
have on each other are dynamic, multifaceted and cross domain. Hence a challenge of
military capability devedopment is the representation and analysis of complex coative
systams so as to mode potential fue environments as oollections of dynanic,
interdependent variables from all domains. Such problem messes require a hierarchy of
modeling techniques, initidly to define the pobem am thn nove tatitiod
Operations analysis approaches providing deeper investigation into specific areas.
The capabilities of the military domain can be rdatively slow to respond to changes in
strategic circumstance, patticulaly when the acquisition of new equipment or changes
to force structure and doctrine are desired, compared to the diplomatic domain which
does not require lengthy acquisition cydes. The identification of what capabilities ae
required to effectively combat the range of possible future scenarios is a process of
belancing the uncertainty of the future with the realities of an economic framework that
necessitates choice between eventualities (Davis 2002, 1). The considerations of these
problems go beyond that of an equipment solution to the full range of capabilities
available in the militay domain and the integration of these capabilities with actions
from the other three domains. Often the strategic planning process can provide guidance
on possible fubre oonflics and illustative plaming somaics, however the
devdopment of the best mix of military capabilities to meet the challenges within these
soenarios is a complex exatise, particilaly when diplonetic, infommation and
economic influences ae altamaive avewes and mot dways ompereniay in
achieving resolution of an issue. For example the Iraq crisis of 2008 demonstrated how
diplomatic negotiations continued with Iraq toward a peaceful resolution on the issue of
‘weapons of mass destruction but the need to preposition military forces for a possible
military eventdlity undemined the sincaity of such diplomatic eforts with the threat
of force.
1.2 Methodology
A method for dealing with the complexities of tansaiing strategic guidance into
capability guidance is militay expeimentaion Ardlyticaly, th tam nilitay
experimentation describes a multimethodology that brings together a range of hard and
soft operations amdysis tedmiques withn a expaimeniaion campaign that
investigates the problen from many perspectives through a series of expaiments and
scenarios (Alberts and Hayes 2002, 43; Curtis and Bowley 2000, 3; Davis 2002, 12). A
desired product from military expaimentation for fore devdopment is the st of
capabilities that perform robustly in both the range of likely scenarios, and the probable
contingencies that arise within those scenarios because of the adversary’s strategy. It is
equally important to be cognisant of the situations in which the set of capabilities ae
insufficient since these represent potential vulnerabilities of the military response that
need to be mitigated through coordination with the other three domains.
An aim of scenario-based analysis for capability devdopment is to identify the aitical
missions that are common across a range of scenarios, and the capahilities required to
perform these missions robustly. The context for the analysis is stimulated through the
conduct of war games that involve not only military players, but also participation from
diponetic and govenment officials. A range of scenarios ae used in the war games
that illustate potential fubre environments, the associated political contexts, and a
vaiely of stressful military missions. The types of war games used are a simple seminar
style where a small number of participants divide into two or more teams and take tums
in plotting moves on a game boad which is then adjudicated by an independent
Qualitative diagrammatic modeling, in the form of influence diagrams (ID), ae used as
an analysis tool to captme the relationships between key variables for each scenario
gamed. The IDs are used in the subsequent design of experiments to test the important
subsystems, defined by military missions and described by capabilities, under a range of
reasonably realistic, but sufficiently different, contexts from those used in the war game.
Figure 1 shows the problem definition process which is aimed at nanowing the scope of
the problem space to be investigated before applying higher fidelity, often more oosily,
analytical techniques, such as moddling and simulation, had opetions analysis or
even further war gaming to investigate specific issues. Once the poblen definition
yprocess has been applied to a range of scenarios a number of focussed expaiments are
conducted, the results of which, in the form of a capability se, are interpreted back up
into the context of the illustrative planning scenarios to demonstrate their significance or
otherwise.
ILLUSTRATIVE SCENARIOS
SCENARIO BASED PLANNING
LULL
WAR GAME RESULTS
diplomatic
APPLICATION
INFLUENCE DIAGRAM
EXPERIMENT SPECIFICATION
FOCUS FOCUS
ISSUE B ISSUE C
Lil
CAPABILITY SET
INTERPRETATION
Figure 1: Problemdefinition process
The application of IDs to military expaimentation is similar to their traditional use as a
lead in step for defining equations and relationships for quantitative system dynamics
modes in that they are used to generale conditions for the design of focussed studies
(Coyle et al 2000, 310; Coyle 1992, 311; Coyle 1983, 885; Coyle 1981, 755). The
focussed study could take the form of a system dynamics model but more than likely in
a military expaimentation campaign it would be just one analytical tool from a broad
vange used to investigate different aspects of the system In a nilitay expaimentation
campaign the ID is a capstone tool capuning the overall system representation that is
then used to specify the conditions for focussed studies using other models. The ID is
then used to situate and relate their results in the broader context of the system and other
taated soenaios with similar system stucte and dements as shown in Figure 2
(Wolstenholme 1988, 10).
CONCEPT FRAME OF REFERENCE
Figure 2: Overarching system representation linked to focussed studies
The remainder of the paper will use a hypothetical scenario to demonstrate the method
used for developing the ID. The tensions in the scenario might require the active use of
military force to a greater or lesser extent or it might be possible to deter the potential
‘aggressor. An ID of the situation is presented which inter relates the political factors,
such as national and multinational support, with the military tasks, such as offensive and
defensive actions. Study of the feedback loops connecting the political and militay
factors show how distinctly different politico-military experiments can be designed to
test the totality of national crisis management capahilities.
2. Hypothetical Political and Military C ontext
The broad scenario description is used to provide participants in the war games a
context for their actions. The soenario described below is representative of a srell scale
regional conflict that, for simplicity, has geo-political constraints limiting the mmber of
actors. The set of military capabilities in the war game can be constrained to a particular
set to evaluate their performance, or unconstrained to explore potential opportunities of
emerging technology and methods of enjployment.
The scenario is set in the near future where over recent years, despite imminent dection
to the United Nations Security Council (UNSC) and the appearanoe of being a good
intemational citizen comprehensively engaged within the region Redland (R) les
stepped up its claims to Northem Mainland of Orangeland (O). These claims have been
supported by a sophisticated infonmation operations campaign aimed at shaping world
opinion on the issue, gaining regional acceptance of Redland’s emaging militay pre
eminence, and orchestrating a Clandestine destabilisation program across all levds of
Govemment in Orangeland.
Redland
Orangelana
Blueland
Figure 3: Redland occupies the northern area of Orangeland and Blueland comes to the
aid of Orangeland
The devastation wought by a recent humanitarian disaster on the north coast of
Orangeland has allowed Redland to initiate its campaign to amex the Northem
Mainland of Orangeland under the guise of humanitarian relief as shown in Rigure 3.
At the request of the Govemment of Orangeland, Blueland (B) will initiate joint military
Plaming to firstly prevent further lodgement of Redland’s forces in Orangdand whilst
tying to pesuade Redland to withdraw its existing foes though diplonaiic,
infonmation and economic actions, and secondly to employ military force to expa them
should they refuse As membership of Redland’s allies on the UNSC is likely to
prevent UN involvement in a regional conflict and as the dlies of Bludad ae
committed dsewhee, the mission will essentidly be a unillde, rather than a
multilateral, operation. This assumption is made to simplify the dynamics of the system
forillustrative purposes only.
Bludand’s military strategic plan has been devedoped as the militay dimension of a
‘Whole of Govemment’ approach designed to re-establish a stable regional seamity
ewironment whilst forestlling direct militay teat to Bludand and its interests.
Although framed by the basic tenet of ‘Strategic Denial’ which is aimed a ooacing
Redland to pull back form Orangdand, the immediate focus is on detening further
Redland expansion into Orangeland, forcing the withdrawal of Redlams nillitay
forces, and restoring Orangedand’s sovereignty through the use of a ooordinated air,
maritime and land campaign Bludand’s nillitary strategic plan has been divided into
the three phases of deter, defeat and devdop. Complete military defeat of Redland
remains the option of last resort because of the rddivdy even maith of nilitay
capabilities. Bludand’s military plamers have been directed by the govemment to
conduct plaming for the military operation to restore Orangdand’s sovereignty in
accordance with the three military strategic phases.
1. Deter Operations to deter further Redland expansion into Orangeland will focus
on demonstrations of Blueland’s forces and an increased military presence in
Orangaland in parld with ongoing diponetic, infometion add economic
pressure.
2. Defeat. This phase is likey to be initiated by a breakdown in diplomatic
negotiations and begin with evacuation operations of approved foreign nationals
from Orangeland. This opportumity will also be used to secure an appropriate
operating bese in Orangeland’s unoccupied tenitory in omer to fadlitate
subsequent operations. The concept of manoeuvre for defeating Redland’s
lodgement in Orangeland is based on three flexible response options - Isolate,
Strike, and Expd - which can be enployed either separately or in combination
in order to force the withdrawal of Redland.
o The ‘Isolate’ option focuses primarily on early air and naval blockade and
interdiction operations between Redland and Orangeland.
o The ‘Stike’ option emphasises coordinated combined strike operations by
air, maritime and special operations dements.
o The ‘Expd’ option is based on land operations with air, maritime and special
Operations to shape the battlespace, and provide support, as required.
The key to defeating any of Redland expansion in Orangeland will be the speed
and effectiveness with which Bludand’s fowes cn oon th initia
lodgement, and cuttail any subsequent force build-up. The lodgement must be
quickly isolated and, if necessary, reduced in strength to achieve a fore
overmatch before ground forces launch the decisive response.
3. Develop. Operations to develop positive post-conflict rdations with Orangdand
and Redland ae besed on engagement though nilitay asistance with
yeconstuuction and nation-building tasks, together with increased exercising and
3. Development of the Influence Diagram
IDs show the causal mechanisms known or believed to operaie in a given system, in this
case the politico- military system of the scenario. They serve several purposes:
1. Putting onto one piece of paper a shared view of a complex problen.
2. The causal influences combine into feedback loops. These may be ether
‘positive’ (reinforcing) loops, or ‘negative’ (limiting or goal seeking) loops. Study
of the loops can provide qualitative insight into the dynamics of the system.
3. The diagram can show the context for other analytical or modelling work.
4. Where necessary and feasible the diagram is the basis from which a filly-
fledged dynamic simulation model can be developed.
The techniques for devdoping IDs of a system have been described dsewhere within
the context of military problems (Coyle 1992, 311; Coyle 1983, 885; Coyle 1981, 755).
An ID for a given problem can usually be drawn in several versions, showing greater or
lesser degrees of defail but retaining conceptual consistency. This section will focus on
how war gaming and the associated processes of military plaming can be used to
generate the ID.
The identification of the key variables, and their dependencies, is the starting point for
the devdopment of an ID using scemaio-besed amlysis. Before binging together
military participants to conduct the planning and wer game it is useful to pase the
documentation that describes the political and military context of the soemario. A
number of candidate factors for the diagram and their relationships to events will be
described in those documents. At this stage it is best to develop lists of these factors, or
use other diagrammatic techniques such as rich pictures or impact wheels, to begin to
form conceptual views of how the factors could be related (Coyle 2008).
Figure 4 shows a possible concept of operations that could be generated by parsing the
guidance given to the commander of Bludand’s military forces for the defeat phase of
the plan. The key dement is how Bludand might conduct an operation that would create
a dilemma for Redland to stay or withdraw from Orangdand. The intent of the concept
of operations is not to achieve success through aititions of Redland’s forces but to
yresent Redland’s commander with a dilemma of aither tisking the forss in
Orangeland being isolated from their mainland and thus rendered less capable, or
withdrawing the forces.
The dilemma in the mind of Redland’s commander is devdoped through which line of
operation, represented by the anows labelled 1, 2 or 3 in Figure 4, Bludad will
manoeuvre and thus how Redland should distribute its forces to combat the possible
lines of operation Bluedand’s tactic is to keep its forces distributed to devdop
inty in the mind of Redland’s commander as to what their objective is. However
if Blueland becomes too aggressive in its disposition towards Redland by messing its
forces at any one location in a concentrated attack, then there is potential for Blueland to
be seen as the aggressor and lose the moral authority to act and thus reduce intemational
support for their operation, after all Redland is in Orangdand under the guise of
humanitarian relief. If Bludand decides to commit forces it must decisivdy counteract
Redland’s reaction while using a diplomatic campaign to belawe the politic
indications of using militay force in oder to maintain the mor athority to at
Bludand’s own dilemma is to balance the amount of fore necessary to threaten
Orangeland sufficiently to force withdraw of Redland, an operational victory, while not
using excessive force which would reduce Blueland’s moral authority and potential lead
to a strategic defeat because of negative wodd opinion Thus by simply parsing the
background material an idealised conceptual view of how the campaign events may play
out can be developed as a basis for development of the ID.
Redland can not
GIEUIT defend both locations
and must move
forces between
ations depending
Redland has split
forces between
two locations
Blueland deploys to
forward operating-be
Blueland
Figure 4: Possible concept of operations for the defeat phase of the campaign
At this stage the variables of the ID have not been devdoped, although the conceptual
elements they relate to and the general factors they will describe have been identified, as
in Figure 4. The next stage is to identify the variables through the observation of the
Planning process used by the participants prior to the war game. The planning process
takes participants through a structured series of deductive and inductive reasoning to
derive the best course of action to take in a situation by analysing factors that relate to
their peroeption of the situation It breaks down analysis of the situation info the
following stages:
a. Study the existing situation
b. Decide the aim to be achieved
c. Considerall pertinent factors and draw deductions
d. Consider possible ways of attaining the aim, and
e. Decide the prefered course of action to achieve the aim.
Courses of action (COA) are produced by the participants prior to the war game using
the planning process. Each COA is described as the series of actions required to achieve
the aim of the campaign and the possible contingency actions that may be required to
counter the likely reaction of the adversary. The COA is a specific instantiation of the
concept of operations represented in Figure 4 described as a series of preferred actions
and possible contingencies.
The indusion of contingency ations, based on the likely reaction of the adversary, is an
important dement in exploring the complexity of the scenario. Without contingencies
the system would be :dativdy simple and deeministic with each side following a
sees of actions imespective of the advasay actions Thus the eglomiion of
contingencies is an dement of the plaming. For example COA 1 may be Bludand’s
refered course but because Redland defected Bludand’s ealy movements Redland
decides to reinforce Bludand’s objective with forces from its other base The detection
of massed forces at the first course’s objective would prompt Blueland’s commander to
take a contingency action that would be to switch to COA 2. The desired effect would
be to place a “wedge” between Redland’s two major concentrations thus effectivey
isolating them from each other in the hope that Redland would then withdraw rather
than face piecemeal destruction. Figure 5 represents such a contingency.
The combination of Blueand’s relative flexibility to change between COAs compared
to Redland’s flexibility in reinforcing either position, and an ability to detect, or better
still anticipate, Redland’s decision to reinforce ae the aiticd factors in Bludand
achieving suocess at the campaign leva. Both of these factors are time dependant and
determined by the capabilities that make up the forces. In going from the representation
of the campaign in Figure 4 to that in Rigue 5 the temporal ad adversary dimensions
have been ackded to the conceptual model of the campaign. However a further dimension
of the information from the planning process is required before the ID can be produced.
if Redland
detects
Blueland’s
‘movements then
Redland is likely
Redland
“Feinforcements then
Blueland would
change to COA 2 to
remaining forges
Figure 5: Course of action showing branch point from COA 1 to COA 2
The final stage in devdoping the ID is to synthesise the conceptual, tempor and
adversarial aspects of the campaign through the war gaming process. It is through war
gaming that the concept of operations is tested and the relationships between key
variables in the COA established. War gaming can be as simple as moving counters on a
board or as complex at using multiple linked simulations The essentid derents ae
representations of friendly forces, a capable and intedligent adversary and representation
of the environment (Peda 1990, 15; Curtis and Bowley 2000, 9).
The influences between variables are a combination of those factors identified during
devdopment of the concept of operations and the decision citeia used by the
participants in the war game. For example Bludand’s moral authority, highlighted in the
concept of operations, will have a reinforcing influence on the suocessful conduct of
Blueland's offensive operations against Redland because it would increase intemetional
support for Blueland’s operations leading to greater range of feasible COA. The
opposite is also true, for example, the planned “Northem Front” in the Iraq aisis of
2008 was limited because of the lack of intemational support from Turkey and Syria
The flexhility of Bludand to change between COA compared to Redlands flexiility
to reanforce either position will also have an impact on Bluedands offensive success.
Thus it is usefil to define a vaiable called Bludand’s opaationl flexihility which is
the combination of the number of feasible COA and the ability to transition efectivedy
between the different COA. The ability to transition betwem COA is influenced
positively by Bludand’s military capability dements such as suitablity of equipment
and taining levds for offensive operations and negativdy by the led Redlands
military capabilities for defensive operations.
Questioning of the war game participants and direct observation of the war game events
is used to capture the relationships of the variables in the ID. The refinement of the ID is
conducted through presenting the diagram back to the participants and obtaining their
feedback on its strucure and content. There is of course not always consensus amongst
participants, particulaly when oonsideing the situalion from the perspectives of
Redland and Bludand in this example so multiple IDs ae accommodated in the
methodology as different Weltanschaumg', or word views, as they are described in
Checkland’s Soft Systems Methodology (Checkland 1984, 166).
3.1 Conventions and notation of the Influence Diagram
IDs can be a little confusing at first sight but, with practice, they can usually be read as
esily as text. The standard conventions of IDs are:
' The Geman expression for worldview. This worldview, or perspective, makes the transformation
process occurring in the influence diagram meaningful. An example of Blueland’s Weltanschauung
would be the sovereignty of nation Orangdand has been threatened and Bludand, as an ally of nation
Orangedand, will take reasonable action to restore that sovereignty,
¢ All vatiables and parameters must be defined so that they could, in principle, be
meesured unambiguously. For example, international perception of Blueland’s ability
to restore Orangeland’s sovereignty Would be unsatisfactory as it is ambiguous,
wherees International support for Blueland's ability to restore Orangeland's sovereignty
is dea: The definition does not imply that the intemetioel community bes
support for Blueland. Were there is no support the variable would be low and if
there were support it would be high Redefining that variable as International
disapproval for Blueland's ability to restore Orangeland's sovereignty woud simply
change the signs discussed below.
¢ A signed anow pointing from variable X to variable Y assets that X has a
causal influence on Y .
« A pilus sign states that, if X increases (decreases), then Y will also incresse
(decrease). In other words, Y changes in the same direction as X. A minus sign
has the converse meaning. Two or more variables can influence Y, and Y can
affect two or more other variables, as required by the problem.
¢ The sign is placed as close as possible to the head of the amow aud as
‘unambiguously as possible where two or more arrows converge.
¢ A lage D denotes a significant time dday in a link If thee ae two or more Ds
in a diagram it is not implied that they ae of identical magnitude and subscripts
can be used if necessary.
¢ Care is taken to minimise the mumber of anows cossing one another but it is
usually inevitable.
¢ Greater care is taken to avoid double- counting the same cause or result.
The special conventions for this example are as follows:
¢ Solid lines ae causal influences within the military context of the campaign
Finely dashed lines are the political and social milieu of the scenario.
¢ Vatiables that can change as time passes in the game are shown in ordinary Arial
type, forexample Blueland moral authority”.
¢ Three ‘output’ variables are emphasised: Intemational support for Blueland to
restore Orangeland's sovereignty, Success of Blueland's offensive operations
against Redland ad Degree of isolation of Redland’s forces. These seem to be
the pimary military phases and political aims of the campaign.
? The changes in typeface correspond to those in the diagram.
¢ Aspects such aS Blueland’s degree of freedom provided by rules of engagement
and Suitability of Blueland’s capability for COA (top of diagram) are so-called
“pressure points’ (PP), or aspects of the problem where investment, changes in
training and doctrine, or political choice, would have an impact on the dynamic
behaviour. Strictly speaking, a pressure point is exogenous to the dynamics of
the system, as is the case with, say, Redland’s level of capability for defence or
Length of time Redland has had to prepare defences. In Figure 6, however, two
of the PPS, Degree of freedom provided by rules of engagement a1] Suitability of
Blueland’s capability for COA ae fattially endogenous in that they ae
influenced by Intemational support for Blueland’s operations .
¢ Firelly, one novel convention is the sign +/0/ (bottom of diagram) to suggest
that Suooess of Bludand's offensive operations against Redland might increase
the Degree of paeived hardship of Orangdand’s population in the region
occupied by Redland, have no effect, or decrease it.
The diagram may seem too broad-bnush but that is largely due to the attempt to get a
representation of the whole of the campaign onto one page, not to represent all the
infonmation from the planning process and war gaming onto a single diagram, as there
are well established products from the planning process that do this. It could be
to show more deal in, for example, Suitability of Bludand’s capability
for COA. That would inevitably add greatly to its complexity and it may be better to
disaggregate segments of the diagram into separate diagrams. For example, Suitability
of Blueland’s capability for COA could be shown as having subordinate pressure points
such as suppression of enemy air defence, battle damage assessment, air task planning
capabilities, and so on. The general me is thal, in a given diagram, the pressure points
should be shown in comparable levels of detail.
For the purposes of the experiment specification process in Figure 1, the degree of
resolution in Figure 6 may be suitable The use of more detailed levels is discussed
bdow for the generation of contexts for focus issue studies.
Blueland’s degree of + Suitability of
freedom provided by rules —_y Blueland’s
fy of engagement capability for COA
t
a +
Intemational support Number of feasible i
for Blueland's offensive— CQA available to Bluelani s ability to
operations Blueland
H € COA
1 ty + Length of time
H Blueland’s operational Redland has had to
Blueland’s moral authority to flexibility nediaiar prepare defences
conduct offensive operations vel of
! = capability
International support -\+ fordefence
for Blueland'to restore Redlands's ability to reinforce
Orahuland's the right objective at the right -
sovepiony me Effectiveness of
“At Redland's prepared
if defences to Blueland's +
i chosen COA
H
i Degree of isolation of
\ Redland’s forces e
. Success of Blueland's
iy ; offensive operations
Redland’s moral authority to againstRedland .
occupy Orangeland and Ability of Redland to .
provide humanitarian support sustain operations i
to the 'ocal population = Orangeland
yt
Degree of perceived
mL hardship of
"“s*=—-Orangeland’s populations“ +/0/=
in the region occupied
by Redland
S.
Figure 6: Influence diagram
4. Generating Experiments
The diagram can support the generation of conditions for subordinate analyses in two
Ways: Via it’s the feedback loops and pressure points.
4.1 Loop Analysis
A feedback loop exists if, starting from any variable and following the anows, one can
retum to the starting point without having passed through any other variable more than
once. It is still a loop if one link crosses over another, so a loop can be snwll and
obvious or very complex and convoluted A given variable can feature in several loops,
as canagiven link (anrow).
From inspection, Figure 6 contains numerous feedback loops, all of which will not be
discussed. For illustrative purposes, two are shown in Figure 7, emphasised by thicker
amrows and identified by descrip tive names.
Both of these loops relate the purely military parts of Figure 7 with the political miliey,
which may well be tie for all the loops in the diagram For these two illustrative loops,
the inner ‘HEAVY HANDED BIG BROTHER’ bop is negative, or resticting, and has
the effect that success of Blueland's offensive operations against Redland is reduced
due to the weakening of Blueland’s moral authority to conduct offensive operations. The
intamational community, while apaoving th dfots by Bludad tw resioe
Orangeland’s sovereignty, as represented by the outer positive reinforcing loop labelled
“LIBERATION OF A SOVEREIGN TERRITORY’, have an avasion to nilitay
offensive action. The toll the opaations have on the local population produces a
conesponding negative influence on Blueland’s moral authority to conduct offensive
operations aS represented by the negetive influence from success of Blueland’s offensive
operations against Redland’.
The relevance of these loops to the specification of expaiments is that they offer two
extemal conditions within which the military must function In one case, the inner loop
dominates, Blueland’s moral authority to conduct offensive operations is diminished, there
Will be litle Intemational support for Blueland's operations ANd Number of feasible COA
available to Blueland Will be limited thus reducing Blueland’s operational flexibility. If the
outer loop dominates, the opposite will apply. Subordinate expaiments should not
attempt to predict which of these conditions dominates, but simply treat them as two
expetimental variants, or hypothesis, used to investigate a set of capabilities.
An interesting vaiaion of this lies in the link to Degree of perceived hardship of
Orangeland’s population in the region occupied by Redland and the sign of 4/04, labelled
“THE CNN EFFECT” in Figure 7. If success of Blueland's offensive operations
against Redland is increasing the plus sign comesponds to a perceived aciverse impact
on the local population through the media because their hardship is increasing resulting
in possibly less local support to Bludland, the 0 means that there is no perceived impact
on the local population, in which case the loop would not exist, and the mimis sgn
amounts to Bludland’s actions reducing the paveived harcship of the local population
through the favourable coverage of Bludand’s actions and possible greater local
cooperation for Blueland’s forces. Three more political oontexts based on local
population support are generated through the +o facility, and no doubt more could be
identified in the other loops.
° There are two reinforcing loops in Figure 7 but concentrate on the innermost through Redland’s ability to
reinforce the right objective at the right time. There are, of course, other reinforcing loops with parallel
paths.
The relevance of the political milieu is dearly stated in the political and military context
eadier in the paper and this preiminary loop analysis has linked those considerations
very Clearly to the dynamics of the military actions through the facilities of the ID.
Blueland’s degree of Suitability of
> freedom provided by rules —_y Blueland’s
fe of engagement capability for COA
H
_# +
International support | umber of feasible Blurland’s ability to
for Biuefand's s* COA available to transition between
Ps Blueland COA
4a,
1 . +H + Length of time
1 ? Blueland’s operational Redland has had to
Blucland's moral authority to flexibility pediand’s prepare defences
conduct of eve operations HANDED EE Gato
" capabili
BROTHER iordefence
Intematiofal support
Redland’s bil ity’to reinforce
for Blueland to restore ; 4
rangelands .,the right obfyctive at the right -
Soren, . Effectiveness of
LIBERATION OF Redland’s prepared
I ORANGELAND defences to Blueland’s
i SOVEREIGN chosen COA
t ‘s.,,, TERRITORY”
i
a Degree of isolation of
‘ A's forces -
\ Success of Blueland's
Redland’s moral authority to nana -
occupy Orangeland and Ability of Redland to =
provide humanitarian support sustain operations in.
to the local -
1 population Omageland
S +
x ;
ag Degree of perceived “THE CNN EFFECT”
hardship of
‘~~ == Orangeland’s population:
in the region occupied
by Redland
Figure 7: Influence diagram with feedback loops highlighted
+H0/~
4.2. Pressure Point Analysis
The pressure points (leveage points is an altemative tam) in Figure 6 have ben
expressed in broad teams both for darity of the diagram and to simplify the specification
of experiments. In the latter sense the notion is to measure, for example, Suitability of
Blueland’s capability for COA a Bluedancd’s capability for offence rdative to that of
Redland’s for defence The relative capability generation idea embraces not only
technologies, per se, such as avionics and missile performance, but also the doctinal,
training and organisational aspects across the full range of capability. Each pressure
point offers another variable to define the experiment variants.
4.3 Specifying Experiments
The loops and pressure points allow the specification of a mission and assocaied
conditions such as military and political constraints and freedoms of action for focussed
studies, as in Table 1. The undedlying principle is to specify expaiments in which it is
not clear what the outcome would be so as to investigate the areas of uncertainty by not
having the campaign deally weighted to any ore side Enploying the pincple of
capability parity between the sides will ensure the expaiments ae infonvetive rather
than being a ‘fait accompli’.
The specification of experiments also supports hypothesis generation with subsequent
focussed studies providing tests of these hypotheses. The hypotheses are based on
Propositions that selected loops dominate system behaviour. Each focussed study of a
hypothesis builds confidence in, or identifies changes to, the overaching campaign ID.
Thus the problem definition process is iterative, with the results of focussed suxdlies
being reflected in an updated campaign ID.
Tt could be predicted that for Experiment One in Table 1, Bludand might be able to do
litte more than deter Redland’s expansion in Orangeland and make a show of force in
the region while being unable to take decisive action because of intemational pressures.
Tn such a cese Blueland would not have the moral authority to proceed past the Deter
phase of the campaign In Expaiment Two, Blueland should be able to restore
Orangeland’s sovereignty through the Isolate, Strike and Expel actions of the Defeat
phase of the campaign The war game indicated that only in favouble politic
Circumstance could such action be possible.
This approach for specifying experiments has several atWvantages, not least its
simplicity, darity, and fledhility. Its many combinations offe a grea vaiey of
experiments may be a mixed blessing. The variety is implicit in the concept of strategic
planning, but the table makes it explicit As with all expaimentation, theefore, carefll
design will be needed to ensure that the expainents are sufficiently diffeent to be
interesting yet not so different that the results camot be compared. It is also important
to be mindful of whether one is using a constrained set of capabilities, in which case
some expatiments will dearly not be feasible and not require further effort, or allowing
new capabilities to be eqored to meet the challenges presented in a range or
experiments. Thus the technique can be used for evaluation of a set of capabilities in
‘various missions and tasks or the exploration of a new set of capabilities.
Tts disadvantage may lie in the capability generation idea, which may seam a little too
broad, as shown in Table 1. An issue is devdoping the most suitable levd of
aggregation so as not getting lost in a welter of detail when specifying experiments. The
disadvantage of the ID technique for experiment specification are the polar requirements
of a high degre of detail for daifying how the capabilities ae geneaied and the
requirement to remain abstracted away from the detail b make the diagrams of use as a
representation of the entire campaign. To disaggregate the IDs into the fine deail of
capability risks the user getting lost in complexity when specifying expaiments. To
avoid this problem a hierarchy of diagrams are used, the most aggregated, such as those
in this paper, represent the entire campaign and the broad issues for focussed studies,
and more disaggregated diagrams are used to support investigations into specific issues.
Table 1: Specifying experiments
Experiment One Experiment Two
Bludand’s Mission o Deer Redland’s] 0 Defeat Redland’s
futher build-up in foroes from.
Orangeland
o Demonstrate o Deloy ad
capability sustain forces
o Increase o Conduct
nilitay stikes against
presence = in Redland’s
Orangdand deployed
forces
o Expd
Redland's
land forces
from
Orangeland
Political o Iner resticive loop] o Outer enforcing
Environment dominates - Bludand loop dominates -
bes litle 3=mod stong support for
authority for Bludand.
offensive action. o Redland remains
o Orangeland’s local defiant.
population is not}/o Orengeand’s local
supportive since they population is
see offensive action supportive of
reducing the ability of Blueland because of
Redland to provide humenitaian
assistance. provided by Bludland.
Operational o Riles of engagement])o ROE mor rdaxed
Consequences vey restictive (eg. (eg. self: defence,
sdf defewe ad interception of any
intaception only of traffic to Orangeland,
Clearly identified ‘unrestricted. action
Redland nilitay against Redland’s
traffic). assds in and around
Orangeland).
available to Bludand| 0 Broad range of COA
reduced to indirect available induding
action. stikes on Redland’s
hondand.
Postulated military o Bheants offersive)o Rediand has a
Capabilities (or capability is able to capability ovennaich
expaimental conditions overcome Redland’s in the land component
from pressure points). defensive capability if but Bludand bes a
1) not capability ovemmaich
reinforced or in natitime and air.
2) Redland o Redad tes hilt
hes ben in significant defences
place for less with the aide and
that 1 month. support of the local
o If or contin is population.
Tt must be remembered that Table 1 is based on consideration of only two loops in
Figure 7, and those loops were so prominently visible as to be obvious. Higue 7
contains an as yet unknown number of additional loops (it might be as many as 10) so it
is certain that other expaimental milieus can be identified and it is very possible that
one ormore of those might specify much more valuable experiments.
Tn addition careful consideration should be given to the issue of whether and, if so, to
what extent, the capability presse points should be disaggregated into ther
components. By keeping the capabilities aggregated the expaiment can explore the
requirements for the capability in general tas rather than specifying the platforms,
structure and doctrine that make up the capability. If the capabilities are disaggregated
into specific platforms, struchwe and doctrine then the performance of that capability
will be evaluated The degree of aggregation is deemined on the intent of the
experimentation, either exploration or evaluation of capability.
4.4 Contexts For Other Studies
One of the uses of IDs is to define contexts in which other focussed studies can be
canied out and to show the relationships between such investigations Figure 7 offas
many examples, the first could be a detailed mode of the effectiveness of Redland’s
defences to cattain COA used by Blueland, with the environment, platform capabilities
and Redland’s ability to reinforce treated in some defail, and using, perhaps, success of
Blueland's offensive operations against Redland @S an outcome measure. Sucha study could
be conducted under the conditions of Expeiment One or Two in Table 1 sesilting in
different test conditions. For example under the oonditions of Expaiment One a
meeting between Redland and Bludand would prove of interest because of
Bludand’s restictive miles of engagement (ROE) and the tactics that would be
employed to combat this. Under the conditions of Experiment 2 the deliberate hatile
may result in a very different outoome because of the impact of more relaxed ROE for
Blueland balanced with better prepared defences for Redland.
The study might be done by any one of a mumber of techniques, such as constructive
simulation or system dynamics moddling. The diagram then places the outoone
mmeesure in relation to other dements of the campaign Figure 6 indicates that success
of Blueland's offensive operations against Redland has a positive influence on Degree of
isolation of Redland’s forces. Thus the results of the focussed study would provide the
conditions under which Bludand’s ability to isolate Redland’s forces is improved or
degraded. Through further qualitative analysis of the diagram the impact the conditions
have on the entire campaign can be identified and thus provide an indication of ther
value not just in the ara of isolation of Redland’s forces but in the campaign ovedl.
The limitation is that only the direction of change of the vaiables is shown though
analysis of the ID not the magnitude or potential nonlineatities in the response that
would fom areas of investment Due to the unquantifiable nate of many of the
vatiables in the political milieu, the devdopment of a comprehensive system dynamics
modd of the entire system is infeasible However qualitative analysis of the entire ID
coupled with quantitative analysis of aitical sub-systems yidds defensble and
auditable insights for military experimentation with an econony of effort.
5. Conclusion
The ID in Rigure 6 has captured the whole of the campaign on a single sheet. To do so
requied an dement of aggregation and that in tum, involved some boadtbush
treatment of capabilities. The result has proven to be useful as a one-page portrayal of
complexity and that may be helpful in its own tight when peseniing the resilts of
military experimentation to inform capability devdopment The amalysis of loops and
yressure points has suggested a concept for the design of experiments to investigate a
range of suitably different politicomilitay conditions under which to eqglore or
evaluate the implications for military and national capabilities.
Theaims of Figure6 were fourfold:
1. The pimary am is to cape key dements of the campaign in one diagram In
addition, the overtly military aspects, such as impact of offensive action, ae
embedded in the milieu of the political aspects, such as moral authority.
2. Putting this complex problem onto one piece of paper has potential as an aide-
memoue for both analysts and participants in the war gaming process. With
suitable explanation, the diagram is a useful quick reference for the extensive
biefing documents that accompany such shudies.
3. The diagram can be used to generale conditions and assumptions for detailed
studies using other analytical techniques.
4. The diagram can be used to qualitatively situate and relate the results of the
efailed studies into the broader context of the diagram.
IDs offer an interface between the high levd strategic guidance provided in illustative
planning scenarios and the detailed opations amalysis trade-off studies of nillitay
capability devdopment All to often the results of later ae left without rdevance
because their purpose and motivation is not dearly linked through to strategic guidance.
In addition, established operations analysis techniques find it difficult to deal with the
unquantifiable nature of political influences that ae dealt with quelitativdy in IDs The
main devdoprent vehide of IDs in this application is war gaming and the associated
military plaming process but the diagrams ae eqally wel suited to a bosiness
exivironment where soenario based analysis is also popular.
References
Alberts, David S., and Hayes, Richard E. 2002. Code of Best Practice for
Experimentation. Command and Control Research Program publication series.
Australian Department of Defence Policy Guidance and Analysis Division 2002.
Future Warfighting Concept. National Capital Printing.
Checkland, Peter. 1984. Systems Thinking, Systems Practice. West Sussex: John Wiley
& Sons Ltd.
Coyle R.G. 2008. Practical Strategy: Structured Tools and Techniques. Pearson
Education, forthcoming November.
Coyle, RG., MJ. Bate, and S. Hamid 2000. Strategic Assessment of Present and
Fube Seauity Environments: A Systems Approach. Defense Analysis Vol. 16, No. 3:
299-314.
Coyle, RG. 1992. The Optimisation of Defence Expenditure. European Journal of
Operations Research. 56: 304-318.
Coyle, R.G. 1983. Who Rules the Waves? A Case Study in System Description. Journal
of the Operational Research Society. 34: 885-898.
Coyle, R.G. 1981. A Model of the Dynamics of the Third World War - An Exercise in
Technology Transfer. Journal of the Operational Research Society. 32: 755-765.
Cuttis, Neville, and Bowley, Dean. 2000. Hierarchical Systems of Enquiry for Analysis
of the Land Force. Proceedings of the Australian Society of Operations Research
Conference.
Davis, Paul. 2002. Analytic Architecture for Capahilities-Based Planning, Mission-
System Analysis and Transformation. Califomia RAND National Defence Institute
publications.
Hood, Robert, and Jackson, Michael. 1991. Creative Problem Solving: Total Systems
Intervention. West Sussex: John Wiley & Sons Ltd.
Groom, Tony. 2002. Tumaround Systems Alignment A Methodology for Dealing with
Complex: Coercive Problems. In Ragsdell et al, eds. Systems Theory and Practice in the
Knowedge Age. New Y ork: Kluwer Academic/Plenum Publishers.
Peda, Peter P. 1990. The art of wargaming: a guide for professionals and hobbyists.
Annapolis, Ma.: Naval Institute Press.
Wolstenholme, EF, 1988. Defence Operations Analysis Using System Dynamics.
European Journal of Operations Research. 34:10-18.
Back to the Top
CC BY-NC-SA 4.0