Share this Post
Advancing Intelligence Analysis for the Multi-Domain Battlefield
Chris Parrett and Tom Pike
In the first paragraph of Carl Von Clausewitz’s seminal work On War he writes …[in war] “more than elsewhere the part and the whole must be always be thought of together (Paret 1976, p. 75).” His words, over a hundred years before the term Complexity theory was coined echoes the common description of complex adaptive systems; “the whole is more than the sum of the parts (Miller and Page 2007).” Complex or non-linear systems have confounded analysis, and often were avoided by science due to their intractability. The increasing speed and veracity of human interaction, fueled by the rapid increase of technological advancements, have only amplified this complexity. Consequently, the complexity in the evolution of war has mirrored in the latest transition from AirLand battle to Multi-Domain battle (U.S. Army Training and Doctrine Command 2017). These advances have also led to the creation of computational tools which provide researchers opportunities to explore and analyze these previously intractable complex systems. The challenge facing the military is how to exploit these computational tools to gain and ensure a competitive advantage. Simply, only computation and its emerging tools provides the ability to analyze complex systems (Axtell 2000). This paper proposes a comprehensive strategy to integrate computational analysis into intelligence analysis and by extension the integration of option exploration tools for Commanders and their staffs to achieve advantage on a battlefield whose dimensionality appears to be increasing exponentially.
Integration of computational analytic tools presents three major challenges (table 1). The first challenge is accepting computational tools to understand and influence complex systems is more than data. In Singapore, arguably the world’s most surveilled society, ubiquitous data collection combined with their Strategic Futures Network, which began in 2009, collects and analyzes data to identify and prevent ‘strategic surprise or ‘black swan’ events. However, in 2011 the ruling party lost 6 seats in the parliament (in their view a stunning defeat) and saw its largest protest in 2013 over a white paper it published on a housing plan (Harris 2014). This unexpected ‘unrest’ occurred as the government formulated decisions and policy based on its massive data collection and scenario development systems implemented to anticipate such things. History is replete with examples of unexpected behaviors which exceed the possible variations data processing could compute. For example, the citizens of Leipzig in 1989 began walking on their Ringstrasse as a protest against an authoritarian state. This protest arose out of seemingly uncorrelated and inconsequential factors (Lohmann 1994). Russians, after the fall of the Soviet Union adapted to the Washington Consensus (a series of policies to convert their economy to capitalism) by creating a virtual economy, an adaption no one imagined or identified until Russia’s economic collapse in the late 1990s (Gaddy and Ickes 2002). The evidence from these experiences argues that an advanced Intelligence Surveillance and Reconnaissance (ISR) enterprise which is processing, exploiting and disseminating petabytes of data may still miss the point. Adaptive systems adapt, and with this the meaning of data collected one day in one location may mean something else another day or in another location. Data collection and processing alone is insufficient, it needs a symbiotic relationship with humans who update and adapt the potential meanings of the data and imagine possible implications. This leads to the second and third challenge of integrating computational analytic tools. How is a symbiotic relationship formed where the computational tools help analysts understand complex phenomena and analysts can explore novel dynamics of complex phenomena in a manner which can still support decisive military operations?
Table 1: Three Challenges of Integration
Dealing with the first half of this relationship, advances in the understanding of complex phenomenon and complex systems occur mostly in universities, institutes and other research groups around the world. Like all researchers on the forefront of developing technologies from ballistics to logistics, researchers seeking to understand complex phenomenon will continue to make advancements. The military’s challenge is how to operationalize these advancements so they can be applied against the practical problems of military operations. Just as the Department of Defense must constantly integrate these new weapons into its inventory, the integration of new computational analytic tools will be a deliberate process required to maintain competitive advantage. For computational analytic tools frequent iterative improvements are a necessity. By establishing these capabilities as services, the tools would be maintained at the service point, providing the software as a service (SaaS), following the same paradigm as many industry SaaS tool update processes, similar to Google Maps or Docs. This provides constant improvement to the analytical tools in a way transparent to the user. In this case, the improvements are emerging theories coded into the software which the analysts and decisionmaker can still understand and employ. This then leads to the third challenge, how are advanced theories represented so analysts and leaders can leverage them to conduct more effective operations?
History demonstrates technological advances only cause a paradigm shift in military operations when the technology is integral to the holistic application of military tools (Boot 2006). Simply providing computational analysis tools is insufficient; the ability to expertly apply the tools must also be instilled. Commanders, analysts, and staffs must adopt a complexity-based understanding of the multi-dimensional battlefield. The scientifically rigorous theories of complexity and the computational support tools must be a fundamental part of Army doctrine.
This will not cause a reconstruction of military doctrine, instead it is the required extension of existing concepts. Clausewitz’s work is rife with concepts that are now the core tenets of complexity theory, not just the previously mentioned the ‘whole is greater than the sum of it parts’. Clausewitz identified the importance interdependencies, in his description of the magnificent trinity where the military, the people and the government are an inseparable interdependent system. In addition, his description of a piece of metal suspended between three magnets is the three body problem, the quintessential non-linear system. (Paret 1976) Military culture has been linked inextricably with complexity theory since before it was recognized as a separate field of scientific study. This allows us to adapt the existing doctrinal framework of Intelligence Preparation of the Battlefield (IPB) into a complexity based framework which can provide a doctrinal handshake with the computational analytic tools.
The development and integration of computational analytic tools is an imperative of the multi-dimensional battlefield. This imperative must overcome three challenges to be successful.
Complexity and IPB
For IPB to deal with the increasing number of domains, it must adapt to a more complexity based framework. Briefly, IPB is the Army’s doctrinal framework for intelligence analysis. It consists of four steps (table 2), (1) Define the Area of Operations, this is a brief description of the area in which a military unit is operating. (2) Describe the Environmental Effects on operations, this step provides a detailed analysis of the terrain and weather and how these factors may impact military operations. (3) Evaluate the threat, analyzes the adversary’s military equipment, their capability, and their doctrinal methods for fighting. (4) Determine threat courses of action, uses the knowledge gained from the previous three steps to develop reasonable hypotheses for how the adversary will likely fight to try and defeat U.S. forces. (Department of the Army Headquarters 2014)
Table 2: Intelligence Preparation of the Battlefield
The challenge with IPB is it does not do well when faced with complex situations. The framework only consists of two entities, friendly and adversary forces, and focuses on the dynamics of one battle at one place and time. This prevents IPB from being able to assess how different domains of even a single adversary force such as cyber, space and ground components may be taking seemingly unrelated action which can have a direct impact on the outcome of a battle in a different section of time and space. This problem compounds when looking at higher echelons who are dealing with more domains over longer time periods. Its shortcomings are also evident when dealing not with one adversary force but with a population who will have multiple entities competing over a variety of interests. In both these cases, Commanders are faced with situations where different combinations of micro interactions may produce vastly different macro effects. This is the essence of complexity, where emergent behavior occurs which is not predicted from the component pieces (Miller and Page 2007; Beinhocker 2006) and ABMs are the primary tool to analyze such systems.
Computational models fall into four categories. (1) System dynamics models, which model the system as an indivisible whole, through stocks and flows processes. (These models are already integrated into Department of Defense planning, but are typically reserved for the small cadre of Operations and Research Systems Analysts (ORSAs) one may find on staff senior staff.) (2) Microsimulations, which model at least two levels, such as the aggregate and household effect of a tax policy change. (3) Discrete event models, which models the interaction of a system’s entities and their associated capacities (e.g. modeling the events from arriving at an airport to plane departure). (4) Multi-agent simulations or agent based models (ABM), which model the individual behavior of heterogenous population of entities or agents, their interaction with the environment and their interaction with other agents.
Currently, researchers and practitioners are employing processes that attempt to holistically address dynamic and complex environment from the top down in order to understand the underlying social system and provide key insights into environmental factors which may affect that system(Davison 2016). While this doctrinal approach aids leaders in understanding the larger geopolitical context, the reliance on intelligence data to provide objective assessments cannot be overlooked. Much of our understanding of complex social systems is derived from our first-hand experience and observation with the components of such systems, specifically, the people, the infrastructures, and local environments. As stated above, however, while we may understand a portion of such system, the interaction of these components amongst each other and their environment often results in an emergent complex behavior. This non-linearity introduces noise and friction into the doctrinal IPB and environmental framing approach.
ABMs are the most appropriate approach for exploring complex phenomenon, as they are the only models which explore how individual situations and behavior can produce unexpected emergent behavior. (Gilbert and Troitzsch 2005). Interestingly, these models share key components with the IPB framework. Both require an in depth understanding of agent- environment interaction. IPB requires in depth analysis of terrain and weather effects analysis to determine how these features will affect adversary behavior. ABMs can use these features as well, and can either program in specifics and/or import GIS data to account for such features.
IPB’s focus is agent-agent interaction with anticipation of how the adversary agent will act and react to friendly actions. Agent-based modeling (ABM) provides a method to disaggregate the various systems feeding the assessments into individual components, which potentially have their own more simple characteristics and rule sets (Heppenstall et al. 2012, p. 85). ABMs focus is also on these dynamics but with more granularity and diversity. Applications using ABM present the ability to model complex issues from the bottom-up, empowering researchers and practitioners to explore complex phenomena and advance analysis beyond the metaphorical or rhetorical examination of the complex systems (Heppenstall et al. 2012, p. 125). It assesses these interactions by programming in specific agent based behaviors and analyzing the results and possibly any emergent behavior. As IPB was intended for combat analysis which is inherently complex and ABMs are the primary tool of complex systems analysis they share key features which facilitates IPB’s adaption into Complex IPB.
Accepting the challenges to integrating computational analytic tools and the use of agent based modeling we propose a strategy with four parts. First, adapt the current IPB framework to ensure it aligns more effectively with complex adaptive systems theory, for simplicity sake we will call this Complex IPB. Second, develop an ABM based on accepted theories of population dynamics and stability, which can be scalable from the village to the national level. This ABM must be (1) consistent with the Complex IPB framework, (2) have the potential for the internal theory to be updated and changed with minimal impact on the user interface and (3) be able to scale data integration from user inputs to large collections of datasets. Third, employ, apply and validate the framework and model in a wide spectrum of stability analysis problems. (The integration of new tools is a challenging undertaking worthy of a separate paper and will not be discussed in detail in this paper.) Fourth, using lessons learned refine and develop a suite of more detailed ABMs analysts across the intelligence enterprise can employ, specifically multi-domain battle. The shortcoming of this strategy is starts with a stability versus multi-domain battle model. This approach, however, is necessary to lay the groundwork of models which can generate the foundational intelligence for more specific models, which are all parts of the same enterprise analyzing different aspects of complex adaptive systems. It is imperative to start this process now to develop the tools required to analyze the multi-dimensional battlefield of 2030 and beyond from the strategic to the tactical level. The first step is to develop a doctrinal framework more aligned with complex adaptive systems theory.
Complex IPB1 is a four-step process, similar to IPB, but shifts the focus of the analysts toward the interdependencies of a foreign system with multiple actors (Table 3).
Table 3: Complex IPB
Step 1: Define the Operational Environment. This step represents an initial assessment of the situation where analysts develop a basic understanding of the different types of groups within the population, and the physical and socio-political environment. For multi- domain battle, this will likely develop to be the domain specific entities within the foreign military organization. Groups can be either integrated with each other in various type of relationships (e.g. symbiotic, exploitative, etc) or separated but still in a common environment with possible interdependencies. The scope and size of the operational environment remains dictated by the command.
Step 2: Describe the Group’s - Environment interaction. For the initial stability model, we borrow from economics and political science. We assume individuals are inherently interested in maximizing their utility (Bueno de Mesquita et al. 2003). The variables selected must directly influence the utility of members within the various groups. The number of variables and the complexity of their interaction can be scaled to fit the allotted time constraints within the decision making process. Through a description of the environment, analysts are inherently developing hypotheses regarding the malleable dynamics of the environment. In this step, analysts gain more fidelity on key revenue sources and governance aspects of the country and how these aspects impact members of the population. Some example questions may entail, what choices do people have to move from job to job or place to place? Are individuals reliant on political patrons? How are grievances over property, trade or other disputes arbitrated?
From an ABM perspective, this step identifies the main agent types and their interaction with their environment. This also provides a main advantage of ABM. Although an analyst may assess five or six different ethnic, political, or class groups in their area of operations and can envision the key features and behavior of each group. Those groups are full of a diverse collection of individuals. By the analyst selecting a distribution of different attributes for each group, the model can replicate a diverse population of agents where under different socio- political conditions individual agents can defect to join an insurgency or other group. The model could also potentially capture the threshold point at which there may be a cascade of defections pushing the society into a revolution or conflict.
This improved ability to identify possible threshold points then gives a glimpse of how the knowledge acquired through developing ABMs like this will provide dramatic improvements when ABMs are developed for multi-domain battle. A common characteristic of complex systems are these threshold points where actions seem to have no effect and then the system tips. As an adversary is a complex adaptive system, a multi-domain version will likely allow analysts to explore at what point a specific adversary domain reaches this threshold and suffers a severe degradation. Only with an ABM could individual domain specific agents constantly adapt to attacks against their domain and could analysts explore where there might be a threshold point to destroy this resiliency. This knowledge would be invaluable to Commanders as they would able to more effectively determine how to employ combat power and have a better approximation of when that combat power will produce the desired effect.
Step two analyzes an entity with a diverse number of parts and how those parts and their environment interact. ABMs are able to extend these diverse parts into diverse agents who compose these groups, providing a level of granularity and exploration cognitively impossible for any human being.
Step 3: Describe Group-Group Interaction. This pertains to how individuals, coalesced into various groups, interact with each other. This step is identical to the wargaming process, analysts use their knowledge to run scenarios on how groups act and react to each other’s actions. In this step analysts, want to envision what actions the different groups may take to maximize their utility. This step becomes particularly hard computationally as agents cannot develop and execute a new and innovative strategy. (Ideally the ABMs we create will allow analysts to input such adaptations based on their imagination.) Even with this limitation the ability for analysts to create and explore a model will provide insights and thoughts they would not have been able to test or research without a model. This process will help them gain better insights into the underlying dynamics which are shaping the population’s behavior. These simulations combined with the analyst’s knowledge will give them a better understanding of when a group will be forced to react and what actions they may take in response. Although the agents may not explicitly develop new strategies, the model will be able to show analysts how individual agents may coalesce into groups or how those groups may dissolve based on the complex interactions of the agents with each other and their environment.
For the multi-domain battlefield, we think this step will allow the commander to envision how different adversary arrangements or disparate parts of the adversary system are working together to impede his or her actions, even if they do not seem correlated. This will also provide the commander a more robust way to understand how specific friendly actions may dramatically undermine adversary effectiveness. Although a significant level of development will be required to reach this level of fidelity.
Step 4: Describe the Possible Scenarios. This step describes how the group interactions will produce the behavior of the population. Different combinations of group behavior as developed in step 3, will produce the emergent behavior of the system. Some combinations may lead the population to civil war, others may cause oppression, while others may force compromise. Analysts can present these possible scenarios to the commander, with a succinct description of what key variables led to these scenarios. The ABM can aid this scenario development by allowing the analysts to test different parameters and observe the range of outputs. We envision this step will be the same for Multi-Domain Battle. In both cases collection, can be used to find evidence supporting the implementation or emergence of one scenario over another.
Complex IPB represents a framework designed to help analysts adopt a complexity based perspective to assessing a diverse foreign population. It shifts from analyzing one adversary, to analyzing the interaction of a multitude of actors. It also provides a framework which captures the basic components of most ABMs, so ABMs can be developed which analysts can employ.
A Prototype Model
Using Complex IPB as a generic framework, we have started to develop an ABM. Although this model uses a variety of theories, it is important to emphasize this is one of many possible approaches. The goal is a computational link between developing theories and analysts. The analysts would continue to use the same or similar interface (e.g. Complex IPB), but the coding would be constantly refined and updated. The following describes the detail of turning theory into any ABM. As this ABM is focused on a stability model as a first step to getting to the level of specificity needed for a multi-domain battle, this description does not reference multi- domain battle.
The prototype model is a stability assessment model with three sequential sub modules, (1) cooperative game theory for Community Based Organization (CBOs) formation, (2) non- cooperative game theory for stakeholder bargaining (this is essentially different CBOs determining if they should form an alliance) and (3) Conflict onset. This model builds on the work of Steven Keptchel who examined coalition formation of autonomous agent (1995), Abdollahian et al. who expanded the coalition formation to group bargaining using non-cooperative game theory (Abdollahian, Zinig, and Nelson 2013), and Yang and Zagorowski who expanded the use of coalition formation and stakeholder bargaining to conflict onset (2017). From our understanding this framework is also consistent with the bargaining model which has dominated the theoretical foundation of the Peace Science Society (Mitchell 2016).
The ABM has citizen agents and one government agent. Each agent has two main attributes of ideology and wealth, which translates into power. Agents react to economic changes that influence their wealth. Changes to their wealth influences their ability to influence the ideological position of their fellow agents. If agents based on wealth, ideology and proximity gain more utility from forming a CBO than not, then the agents form a CBO. This then leads to the stakeholder bargaining module. In this module, CBOs engage in rounds of negotiations where all parties attempt to maximize their interests with respect to their ideological position. Challenges to the government occur when CBOs differ ideologically from government above some threshold and the CBOs feel they have the power to challenge the government. If the CBOs do challenge the government then a conflict occurs and the model terminates.
When the model is initialized, agents are assigned initial attributes of location, ideology, selectorate membership, and wealth. These attributes are updated at each iteration of the model as their preferences and wealth change. The model then transitions to the first module. Agents form a pairwise CBO with other agents based a distance input variable. These two agents then evaluate the utility of this partnership using calculations based on their ideological preference and wealth (see diagram for specifics). An agent compares his utility of staying alone and the utility of participating in a CBO with one other agent. The agent decision process uses Bilateral Shapley Values from cooperative game theory, which splits the benefit of forming an organization based on each agent’s contribution to that organization (Ketchpel 1995). After a citizen agent joins a CBO, which consists of two agents, their preference changes to the CBO preference. Each agent forms at most one CBO at each iteration, and these CBOs are placed in a list which contains all CBOs the specific agent has formed. At each time step, after determining to form a new CBO, the agent will also review all its CBOs and determine whether to maintain those CBOs based on updated preference and wealth due to other CBO formations. After successively more iterations agents will constantly review their list of CBOs and dissolve any links which are no longer beneficial based on their changing preferences. When the agent has numerous CBOs, its preference is updated to the preference of the CBO least different from its current preference and the maximum wealth of all its CBOs. After doing this for all agents the model then transitions to the next module.
Diagram 1: Flow Diagram of Modules of Prototype ABM
In stakeholders bargain module, we use non-cooperative bargaining model to reflect competing interests during the process. CBOs formed in the first module decide whether to stay alone or get together with other CBOs to form larger coalitions. Similar to the CBO module, after each stakeholder agent calculates their utility and makes a decision in the first round, coalitions form with preferences weighted by power and a subsequent power is calculated as the sum of the coalition members. Then each individual stakeholder and each citizen agent in CBOs participating in the stakeholder module, updates their preference according to the coalition preference. Coalitions then calculate their utility based on their new preferences to decide if additional utility gains can be achieved by joining a new coalition. All coalitions and stakeholder agents perform the same process each round constantly forming and breaking links with each other based on how the changing preferences of the population is shaping and being shaped by their preferences.
In the third module, the stakeholders (CBO coalitions) decide whether to challenge the government. If the CBOs coalition’s power is greater than a predetermined threshold (inputted by the user) and the ideological distance between the CBOs coalition ideology and the government ideology is sufficiently large then they challenge the government.
Although this description may seem far from the Complex IPB framework, it has the key aspects of the framework. This process represents steps 2 and 3 of the process. Agents interact, form coalitions based on their ideology and wealth, from this formation they examine their situation and if they have the power and means they challenge the government. This disparity reinforces the need for such models as they will help the military integrate new theories into existing frameworks which most analysts do not have the time to monitor, learn or integrate. This prototype needs to undergo more validation to ensure it matches dynamics seen in the real world. It is, however, based on widely used and accepted theories from political and conflict science (Mitchell 2016; Bueno de Mesquita et al. 2003; Abdollahian, Zinig, and Nelson 2013). We currently have different prototypes, but they require more development. After more development, an interface will be placed on the model which allows the analyst to input characteristics of the environment, and specific groups with their ideology and wealth attributes, in a manner consistent with the Complex IPB framework. Our goal is to have model which imports some of these attributes, such as GIS data for the environment, or demographic information for different groups and their respective wealth as part of the overall economy. Our timeline to accomplish this is the next 18 to 24 months.
The emergence of the multi-domain battlefield is indicative of the increasing complexity of warfare. The integration of computational analytic tools is a necessity to visualize and understand this complexity. Effective tools must be about more than data, they must enable analysts by providing understanding to the petabytes of data being collected. Analysts must be able to alter how the data is organized, understood and collected so they can find meaning in unknown and unexpected ways. Critical to this ability is the analysts’ creation and implementation of ABMs, which are based on rigorously developed theory, to explore their phenomenon or battlefield of interest. For these ABMs and the analysts use of them to be effective they must be intimately tied to military doctrine, so they can be understood and exploited by commanders and their staffs. To aid this process we proposed the further development of IPB to Complex IPB. This will allow analysts to deal with more complex situations by having them understand the foreign system in terms of the diverse and adaptive entities which compose them and mirrors the fundamental components of ABMs. We then outlined a working prototype stability assessment model. We used a stability model, because we do not assess the technology yet exists to get to the specificity required for a multi-domain battle. A stability model provides two advantages to achieve this goal. It will help develop and integrate ABMs, which generate lessons and understanding to develops models which handles greater levels of specificity. These models will also provide the foundational intelligence necessary to build specific multi-domain battlefield models. Computational analytic tools are a necessity for the multi-domain battlefield and it is essential to begin their development now.
Abdollahian, Mark, Yang Zinig, and Hal Nelson. 2013. “Techno-Social Energy Infrastructure Siting : Sustainable Energy Modeling Programming ( SEMPro ).” Journal of Artifical Socieities and Simulation 16 (3): 1–12.
Axtell, Robert. 2000. “Why Agents?: On the Varied Motivations for Agent Computing in the Social Sciences.” Center on Social and Economics Dynamics - The Brookings Institution, no. 17: 1–23. doi:10.1016/j.cep.2007.02.029.
Beinhocker, Eric D. 2006. The Origin of Wealth. Boston, Massachusetts: Harvard Business School Press.
Boot, Max. 2006. War Made New: Technology, Warfare, and the Course of History 1500 to Today. New York: Gotham.
Bueno de Mesquita, Bruce, Alastair Smith, Randolph M. Siverson, and James D. Morrow. 2003. “The Logic of Politcal Survival.” The MIT Press.
Davison, Ketti. 2016. “Environmental Framing: Understanding a Complex World.” Military Intelligence Professional Bulletin 42 (3): 59–61.
Department of the Army Headquarters. 2014. “ATP 2-01.3 Intelligence Prepration of the Battlefield.”
Gaddy, Clifford G., and Barry W. Ickes. 2002. Russia’s Virtual Economy. Washington D.C.: Brooking Institution Press.
Heppenstall, Alison J J, Andrew T. Crooks, Linda M. See, and Michael Batty. 2012. Agent- Based Models of Geographical Systems. Agent-Based Models of Geographical Systems. Springer Netherlands.
Kauffman, Stuart. 2016. “Keynote Address: International Congress on Agent Computing.” George Mason University.
Ketchpel, S P. 1995. “Coalition Formation among Autonomous Agents.” From Reaction to Cognition, no. 957: 73–88.
Lohmann, Susanne. 1994. “The Dynamics of Informational Cascades : The Monday Demonstrations in Leipzig , East Germany.” World Politics 47 (1): 42–101. http://www.jstor.org/stable/2950679.
Miller, John H, and Scott E Page. 2007. Complex Adaptive Systems: An Introduction to Computational Models of Social Life. Princeton Studies in Complexity. Vol. 27. doi:10.1016/S1460-1567(08)10011-3.
Mitchell, S. M. 2016. “Dangerous Bargains with the Devil? Incorporating New Approaches in Peace Science for the Study of War.” Conflict Management and Peace Science 34 (1): 1– 19. doi:10.1177/0738894216667621.
Morris, Victor R. 2017. “Complex Intelligence Preparation of the Battlefield in Ukrainian Antiterrorism Operations.” Military Review January-Fe (February): 58.
Paret, Michael Howard and Peter. 1976. “Carl Von Clausewitz, ‘On War,.’” What Is War, Ed. and Trans., 83-101-737.
Pike, Thomas, and Piotr Zagorowski. 2016. “Dense Urban Areas: The Case for Complex IPB.”
Military Intelligence Professional Bulletin 34-16–3: 28–36.
Pike, Tom, and Eddie Brown. 2016. “Complex IPB.” Small Wars Journal.
Pike, Tom, and Eddie J Brown. 2011. “Populations as Complex Adaptive Systems: A Case Study of Corruption in Afghanistan.” Small Wars Journal. smallwarsjournal.com.
Pike, Tom, Nick Long, and Perry Alexander. 2015. “Understanding Nations: New Ideas to Analyze Foreign States.” InterAgency Journal 6 (3): 47–56.
Yang, Zining, and Piotr M Zagorowski. 2017. “Climate Variability , Opposition Group Formation and Conflict Onset.”
1. This is a further refinement of previous complexity based analytic frameworks. In this version, the number of steps is reduced to four. (Tom Pike and Brown 2011; Tom Pike, Long, and Alexander 2015; Tom Pike and Brown 2016; Thomas Pike and Zagorowski 2016; Morris 2017)