2021 Expressions of Interest – call for Operations Research (OR) Collaboration
DSTG undertakes rigorous, scientifically based analysis of Defence operations and capabilities to provide independent, impartial and timely advice.
Our Mission is to develop and employ trusted analytical methods and decision support tools that give Defence and national security decision superiority across all aspects of force design, operational planning, command and control, and support to the Australian Defence Force on operations. DSTG and the state-based innovation networks are launching a call to academia to collaborate on a number of Defence OR topics through the Operations Research Network (ORnet).
The intention is that projects will run for a period of up to 12 months commencing in July 2021. Individual project proposals can indicate work up to $100K. UPDATE: In-kind contributions are welcome, however there is no expectation of their inclusion in a successful proposal.
Submissions are now closed.
The ORnet is a collaborative initiative between Defence and academia that will bring together a network of universities to work on Defence problems. It was established to grow and strengthen the Australian operations research (OR) community’s ability to support Defence decision making, with a particular focus on Australian Defence Force design and employment. The intent of the ORnet is to provide coordinated pathways by which Defence can create enduring relationships between DSTG and academia to address challenging OR problems of national and strategic importance.
DSTG is seeking to revolutionise how we undertake OR to better handle the interaction of complex geopolitical, social, technological, economic and cultural factors impacting the capability of the current force and design of the future force.
Topics of Interest
This call will be comprised of two components:
- Philosophy of OR
- General Defence OR topics.
Universities who are signatories to the Defence Science Partnerships (DSP) 2.0 program are welcome to submit proposals to any/all topics contained in this call, and the Commonwealth reserves the right to fund all, some or none of the topic areas.
Philosophy of OR Topics
1. The Epistemic Challenge: Coping with Uncertainty, Complexity and Indeterminism
The epistemic challenge arises from the mismatch between the levels of certainty sought by decision-makers, the suite of formal techniques available to OR practitioners, and the nature of the most important planning and decision problems facing society.
Defence seeks to establish OR (and the broader science of decision-making) as the pre-eminent science for tackling complex, strategic and interconnected problems that are characterised by fundamental uncertainty, indeterminism and self-referentiality. At the strategic level uncertainty manifests in the limits to our knowledge of complex and changing decision problems subject to the actions of indeterministic exogenous influences, typically of human origin. Traditional OR methods are unable to adequately account for the inherent unpredictability of such contexts because they presuppose a degree of stationarity whereby future statistics remain the same as those of the past. This has led to the situation whereby practitioners concern themselves chiefly with the ‘concrete’ features of their decision problems, typically at a more functional level, and have avoided contexts featuring fundamental uncertainty.
In contrast, we seek to employ OR on complex, strategic and interconnected decision problems and therefore seek a new epistemological framework to guide methodology development in situations of fundamental uncertainty. The new framework will eschew the assumption that posits that limits on knowing represents the proper limits of applied science and prevent robust decision-making. Rather we seek a new epistemology that critically grapples with uncertainty and indeterminism whilst still providing for robust and effective decision making. Central to the project will be the attempt to deal critically with the challenges associated with bounding conditions on the uncertainty of properties of interest for decision-makers. These bounds capture information about what is knowable and unknowable within the decision problems and its environment. It is anticipated that the role of the practitioner in these contexts would be to determine the most effective bounding conditions to utilise in choosing courses of action designed to reduce surprise.
2. The Representational Challenge: Coping with Complexity and Interconnectedness
The representational challenge arises from attempts to bound decision problems of interest that are fundamentally distortive to wider (and potentially more important) problem contexts. It produces the phenomenon known as ‘sub-optimisation’ within OR.
Defence seeks to establish OR (and the broader science of decision making) as the pre-eminent science for tackling complex, strategic and interconnected problems that have, thus far, defied mathematical representation. At the strategic level decision-makers are confronted with complex webs of interconnected problems that impact upon each other in unpredictable ways. The context is characterised by huge numbers of potential variables and relations amongst them as well as connections to a vast array of additional problems. In such an environment mathematical representation has proved elusive, either because of the nature of our current knowledge, the nature of the phenomenon, or both. Because the dominant epistemology within OR seeks a mathematical representation of the systems that it studies, the discipline has, over time, become associated with more functional problems that have relatively low complexity and therefore unambiguous representations and solutions.
In contrast, we seek to employ OR on complex, strategic and interconnected decision problems and therefore seek a new epistemology from that of mathematical representationalism. The new epistemology will eschew the dominant ‘predict-and-prepare’ paradigm that sees the development and optimisation of representative models as the raison d’etre of OR. Rather it will seek a form of representationalism that critically grapples with the fact that models of complex decision problems have justification break-off points with respect to original problem boundaries and these may be distortive to wider ‘exogenous’ contexts that may be unknown and potentially more important to both clients and society writ large. Central to the project will be the attempt to deal critically with the challenges associated with defining systems/problems of interest and the presence of unknown, exogenous, unpredictable and changing variables and relations. It is anticipated that the new epistemology will posture Defence to better anticipate 2nd and 3rd order impacts associated with strategic decision problems and avoid adverse unanticipated consequences linked to the narrow problem boundaries.
3. The Teleological Challenge: Coping with Subjectivity and Purpose
The teleological challenge arises from attempts to justify the normative and axiological components of planning and decision-making.
Defence seeks to establish OR (and the broader science of decision making) as the pre-eminent science for tackling complex, strategic and interconnected problems that are permeated with normative, teleological, axiological and ethical considerations. At the strategic level decision-makers are confronted with complex decisions about imperatives, priorities and impacts that have the potential to affect a wide variety of stakeholder groups, some of which may not be involved in the decision process. Policy proposals are invariably relative to the original problem framing and therefore the contingent boundary judgements imposed upon the world by OR practitioners and their clients. Indeed normative judgements about imperatives and implications, costs and risks, values and criteria infuse all aspects of OR practice at the strategic level. This introduces a relative and subjective aspect to OR practice because different stakeholders are motivated by different imperatives which leads to different notions of improvement. Due to the fact that OR sees itself as a neutral, value-free, objective and empirically verifiable science the discipline has largely limited itself to the instrumental considerations of ‘means’ (as opposed normative considerations linked to ‘ends’). This has in-turn driven a focus on functional problems rather than strategic ones.
In contrast, we seek to employ OR on complex, strategic and interconnected decision problems and therefore seek new ethics for the discipline. The new ethical framework will eschew the dominant ‘given-the-problem’ school of thought that sees practitioners refer all teleological and axiological considerations to clients and concern themselves solely with finding the optimal means for achieving given ends. Rather Defence seeks an ethical framework that critically grapples with the fact that all recommendations from OR practice represents somebody’s solution to somebody’s problem and hence have a normative content in need of legitimation. Central to the project will be the attempt to deal critically with the challenges associated with subjectivity, purpose and power dynamics. It is anticipated that the new ethics of OR practice will posture Defence to better consider the organisational, social, political and ethical impacts associated with strategic decision problems and avoid the charge of advocacy analysis linked to the uncritical acceptance of the norms, values and purposes that infuse strategic decision problems.
4. The Meta-Methodological Challenge: Coping with Paradigmatic Incommensurability
The meta-methodological challenge arises from attempts to develop a comprehensive understanding of important decision problems by finding complementarities across disciplines and unifying these within a pluralistic meta-framework.
Defence seeks to establish OR (and the broader science of decision making) as the pre-eminent science for tackling complex, strategic and interconnected problems that defy comprehensive framing from within the methodological imperatives of a single discipline. At the strategic level decision-makers are confronted with complex decisions that require insights from numerous disciplines, each with their own ontological, epistemological, methodological and linguistic frameworks. However, we have no vantage point from which we can understand the complex relations between these disciplines and attempts to do so regularly underestimate the differences. Indeed the intellectual enterprise is characterised by constant flux in which overlappings appear at the edges of disciplines and, from these, new schools of thought are created, deconstructed and thence re-contextualised again and again. Moreover, meanings are changed from one context to the next and mutual understanding is often elusive. When faced with these challenges OR practitioners have tended to restrict their attention to more functional problems that can be adequately addressed from within a single ‘quantitative’ framework.
In contrast, we seek to employ OR on complex, strategic and interconnected decision problems and therefore seek a new meta-methodology for the discipline. The new meta-methodology will eschew approaches that pre-suppose a ‘god’s eye view’ above the disciplines, dissolve into an imperialistic ‘reconciliation under duress’ or rely on ill-defined (and gender-laden) dichotomies such as ‘hard’ vs ‘soft’. Rather Defence seeks a meta-methodological framework for OR that critically grapples with the need for insights from across the disciplines whilst simultaneously avoiding totalising frameworks/discourses and retaining the distinctiveness of each area. Central to the project will be the attempt to deal critically with the challenges associated with the fetishisation, reification and hegemony of dominant paradigms and the creation of legitimate spaces for incommensurate and otherwise dissonant approaches. It is anticipated that the new meta-methodological framework for OR will posture Defence to better understand the multifarious nature of complex strategic decision problems and identify paradigm-busting advancements that emerge from the interfaces between relevant disciplines.
Defence OR Topics
1. Visual Representation of Military Concepts using Virtual Environments
The representation of military concepts often takes the form of simple graphics and supporting text, which is inadequate for representing future complex operating environments or large complex capability systems with many dynamic elements. Consequently, there is a need for tools that enhance immersion and enable visualisation of new concepts, technologies and capability systems. We have recently explored the use of Virtual Reality (Unreal Engine/Infinite Studio and Occulus Rift) for this purpose. The project was successful as a ‘proof of concept’, but further work is needed to more rapidly generate and refine VR products from initial ideas, and incorporate richer tactical behaviour in the concepts. There are two, related projects that build on this work.
The first project would use VR to demonstrate how emerging technologies could be combined with future military systems and demonstrate the effects that could be generated (e.g., a high energy laser on an uninhabited ground vehicle). These artefacts would be used as an input to subsequent activities to design novel concepts for how military forces could employ the technologies. The aim is to improve the design process by using immersion to drive creativity and enable collaboration, communication and common understanding.
The second project would use VR to represent these concepts dynamically, showing the tactical employment of the technologies by the military force. This would be used to better communicate the concept to stakeholders, complementing simple graphics and supporting text, and helping reduce different interpretations of the concept which could arise.
2. New Methods for Forecasting Technology Convergence Points and Capability Effects
Technology foresight provides evidence-based decision support to strategic planners by forecasting potential impacts of new technological developments. While technology foresight research has largely been focussed on understanding and forecasting the implications of specific technologies in their own right, new foresight methodologies are required that determine how different technologies and tactics converge to deliver disruptive capability for the military & national security domain.
The scope for pursuing research that may address this challenge is broad and example research foci may include:
- Advanced foresight methodologies to support analysis and discovery of novel opportunities presented by emerging technologies in addition to addressing known capability gaps.
- Development of a systems-based view that defines capability as the capacity to deliver a desired effect. This systems-based view should aid Defence in discovering how convergence of technologies may deliver new effects.
- Development of methods to identify the opportunities presented by emerging technologies from the adversary perspective and potential operational models (red-teaming).
This research is intended to deliver the following:
- A methodological framework (e.g. a set of guiding principles and steps) for development of operational models that leverage emerging technologies and for elicitation of capability effects;
- of methods, tools and techniques that can be used within the methodological framework (this may include templates, software, remote-based options, indicative case studies);
- Validation frameworks for employing the selected methodology and methods (e.g. What ensures robustness of results? What is the scope for data triangulation? What is the Method Readiness Level?)
- Tools for iterative refinement and knowledge transfer (e.g. manuals, workshop scripts, ScriptsMap, etc.)
3. Comprehensive Simulation Meta-Modelling Capability
One of the key analytical activities in simulation analysis is the fitting of a ‘white-box’ meta-model approximation to a ‘black-box’ simulation through a design of experiments (DoE). Land Capability Analysis (LCA) branch’s meta-modelling currently relates individual outputs (at simulation termination) with input factors.
Suppose you run two simulation experiments. One with 5 Blue tanks and one with 10 Blue tanks (while all other inputs are held constant). At each simulation termination, a metric is recoded indicating whether Blue won or lost. Due to the stochastic nature of the simulation 100 replications of each experiment were performed.
Suppose the simulation data resulting from this experiment was Blue won 50 times with 5 Blue tanks and 80 times with 10 Blue tanks. From this data, one could test statistically the significance of the difference of the two success proportions and/or fit a regression model to quantify the effect on the probability of Blue success that changing the initial number of Blue tanks is predicted to have.
Extending this simple one-factor experiment to multiple inputs and their interactions via an appropriate DoE and multi-variate generalised regression modelling is a standard and effective way of characterizing the approximate functional relationship between simulation inputs and the value at termination of a simulation output of interest.
However, while this input-output regression analysis can answer ‘what’ questions, like ‘what is the expected increase in the probability of Blue success if the initial number of Blue tanks is increased by 5?’ it can say nothing about ‘how or why’ that effect occurs.
Most simulations record data on a large number of internal variables that change over (simulated) time. Necessarily this includes the initial values of the input factors and the terminating value of the metric of interest in the above design of experiment and regression analysis. But conceivably, every possible variable describing the complete state-space of the simulation could be captured. Our interest is in how this internal data can be exploited in answering the ‘how or why’ questions behind the ‘what’ answers of regression.
For the 50 replications where Blue won with 5 Blue tanks, is it possible to ‘analyse’ the associated internal data and ‘partition’ the 50 replications into a small number of subsets, each of which is ‘characteristic’ of a ‘different’ but ‘meaningful’ ‘path’ of progressing through the simulation to Blue success?
Suppose we could do likewise for the 80 replications where Blue won but with 10 Blue tanks. The idea is that a meaningful ‘comparison’ of the two sets of ‘paths’ to Blue success might suggest an ‘explanation’ of how or why the increase in the initial number of Blue tanks affected the probability of Blue success to the extent it did.
Under this Collaborative Research Project, LCA wishes to partner with University academics to develop a more comprehensive simulation meta-modelling capability, which can exploit these output metric correlations and internal simulation data.
4. Aggregation Level for Australian Campaign Simulation Tools
Australian Joint Force development requires a suite of simulation models to address questions at different levels of fidelity. This suite of simulations requires different levels of aggregation and because of the depth and breadth of the questions being addressed by them. For example, in providing evidence to support tactical land combat decision the simulations are fundamentally different to those required to support theatre level land combat decision. Domains, such as land, maritime and air, require different simulations because the features important in land tactical combat simulation are not necessarily present in maritime or air tactical combat simulation. When developing a suite of campaign level simulation tools for integrated joint (across all domain) force development it is necessary to use the right level of aggregation across these tools so that their results can be consolidated. This is likely to be a different level of aggregation to our larger allies simply because we have a smaller force. We are looking for an academic partner who can identify an evidence-based process to decide on the correct level of aggregation in each of the domains when looking at whole of ADF questions and work with us to implement that process.
5. Innovative Design
We are interested in academic partners who will work with us to develop innovative military-focused design approaches that will enable the development of force design options (organisational structures and capability systems) that will facilitate future operational success by the ADF. To design a robust and adaptive ADF we are looking for design approaches that generate options that are adaptive under disruptive innovation and technology change. The process must also ensure there is a key evidence base for any alternative design paradigms that are generated. We seek designs that are highly capable in the sets of scenarios we envisage spanning the space of potential futures but which can adapt to futures that were unforeseen in our scenarios. Ideally, the methodology developed will produce designs that disrupt our potential adversaries, creating surprise and imposing significant costs to counter.
6. An Extensible Framework to Compare Operations Research Methods
Australian Joint Force development requires analytical campaigns where multiple methods are used to develop a nuanced understanding of the problem and solution spaces, leading to appropriate coverage and allowing triangulation. We seek a general and extensible framework for the applicability of methods outlining under what circumstances each method would be used, situations in which particular multiple methods should be used together and articulation of the additional information provided by using specific sets of methods. In this project we seek the development or identification of an extensible framework for comparison and its application to the comparison of Qualitative Bayesian Belief Networks and the Delphi Process.
7. Understanding Future Robustness through Feasible Scenario Spaces
Understanding future robustness through Feasible Scenario Spaces has been proposed in Bowden et al. “Feasible Scenario Spaces: a new way of measuring capability impacts”. Feasible Scenario Spaces identify the futures that a given force option is able to address with acceptable risk. This and other measures of future robustness are critical to understanding the ability to deal with the irreducibly uncertain future for which Defence must prepare. These theoretical approaches need to be operationalised to allow them to be applied in practice. Operationalisation may require coupling of the Feasible Scenario Space approach to a method for systematic exploration of the parameter space such as General Morphological Analysis. We are interested in working with an academic partner to understand how we can measure the ability of a force option to deal with possible futures.
Call for proposals
Interested Australian universities are encouraged to identify relevant fields of endeavour or expertise where they would be willing to engage and partner with DSTG. Universities are requested to submit proposals in response to the OR topic of interest. Proposals should include
- the topic title(s)
- how they heard about this ORnet call for collaboration
- research methods proposed
- an outline of the research outcomes relevant to the topic(s)
- relevant research experience and papers
- a costing estimate.
Proposals should be a maximum of 2 pages. You are encouraged to use the template available in the Attached Files to format and articulate your proposal.
The selected applicants will then be invited to engage with the DSTG topic owners, to discuss and potentially develop a more detailed research program.
Where appropriate, there will be an opportunity to aggregate multiple proposals into multi-party agreements at a later time, up to a maximum total value of $200k. DSTG will continue to play a central role in this relationship, but academic organisations may also wish to form their own connections and propose multi-party research teams to DSTG.
Monday 1 March 2021 – Call for research collaboration proposals
Friday 26 March 2021 11:59 PM – Closing date for research collaboration proposal submissions
From Wednesday 7 April 2021 – Selected partners are notified and invited to begin discussions on details of the collaboration
Monday 12 April to Monday 31 May 2021 – Discussions about work program requirements with selected partners
Tuesday 1 June to Friday 2 July 2021 – Establishment and executions of collaborative research agreements
Monday 5 July 2021 to Thursday 30 June 2022 – Duration of collaborative research agreements
Submitting Expressions of Interest
To submit an EOI for one or more research topics, or if you have any queries about the process or topics, please e-mail BEargPbyynobengvba@qfg.qrsrapr.tbi.nh.