Publication

The purpose of the Enterprise Systems Analysis line of research is to develop and evaluate a methodology for modeling and analyzing enterprise systems. We define an enterprise system as a set of interacting organizations that serve a purpose yet have no locus of control. Their behavior is often complex and must be viewed simultaneously from several different perspectives to be understood. The US Department of Defense (DoD) faces a number of challenges where there are multiple interacting organizations with no central locus of control. For example:

• Combating the proliferation of counterfeit parts in military systems
• Managing joint and international acquisition programs
• Coordinating disaster and humanitarian responses involving governments, NGOs, and US agencies
• Sustaining the defense supplier base in the face of declining acquisition quantities

Consequently, DoD has requested research to enable DoD and Government policy makers to better understand these enterprise problems and shape policy appropriately. More specifically, any enterprise systems analysis methodology should enable:

• Representing the “as-is” enterprise, the “to-be” enterprise, and the path between them
• Understanding relationships between variables and techniques for projecting outcomes and performance
• Providing a means for experimentation and creation of response surfaces for analysis of key tradeoffs
• Providing a systematic method to search for policy tipping points and identify counterintuitive results
• Creating an interactive environment for discussion and debate of strategies, policies and plans
• Enabling key stakeholders to understand the implications and potential second order effects of policy and resource decisions

The work performed during this research task (RT-161) is direct follow-on to the work performed during RT-138 and RT-110. The outcome from the prior work was a shift in emphasis away from building a unitary enterprise model toward a core-peripheral approach in which “peripheral” models could be added or removed as needed to generate scenarios of interest to enterprise stakeholders. Also highlighted, via a series of peer-reviews, was that the methodology needed to be enhanced to better detect unintended or counter-intuitive policy consequences and to better deal with multi-scale ontologies. Consequently, the major tasks for RT-161 were:

1. Apply the core-peripheral approach to a case study of protecting critical infrastructure (Section 0)
2. Develop and validate counter-intuitive results, secondary effects, and policy tipping points (Sections 0 and 0)
3. Extended canonical phenomena and model reuse methods to include multi-scale ontologies (Sections 0, 0, and 0)
4. Update the enterprise analysis methods to incorporate the results of the other tasks (Section 0)

From the execution of these tasks, we were able to develop a fairly substantial update to the enterprise modeling methodology. More specifically, we found from the application of the core-peripheral approach to the critical infrastructure case study that the peer reviewers were interested in using the model for analysis and insight. This is in contrast to the results of the counterfeit parts case study (RT-138, RT-110) where the peer reviewers tended to focus on the use of the model for communication. While this is by no means an absolute validation of the core-peripheral approach, it is an encouraging result. Beyond the case study, a theoretical investigation yielded insights on to how to partition an enterprise system across multi-scale ontologies to generate the core and peripheral models as well as how they should be used together to detect the unintended consequences of a policy. Ultimately, this lead to the revision of the enterprise modeling methodology that reorganized the ten-steps into three major phases. Each phase contains a number of more detailed steps that should provide additional guidance to enterprise analysts. Finally, we also identified a number of promising avenues for future research to better improve the efficacy and applicability of the enterprise modeling approach.

The remainder of this report is organized as follows: Section 0 briefly reviews the findings of RT138 and RT-110 to explain and motivate the work performed during RT-161. Section 0 presents the results of applying the core-peripheral approach to a case study of critical infrastructure protection. Section 0 summarized the results of an Industry-Government workshop held to discuss the challenge of model centric-engineering approaches which share the same technical and organizational challenges as model-based enterprise analysis approaches. Section 0 provides a detailed literature review of how multi-scale ontologies are modeled and how counter-intuitive results are detected in both the physical and social sciences. With regard to the multi-scale ontology aspects of the problem, Section 0 develops a detailed mathematical analysis of the problem to suggest necessary conditions as well as approaches to mitigate the challenges of modeling across multiple scales. With regard to detecting counter-intuitive results, unintended consequences, and policy tipping points, Section 0 develops a proposed approach to partitioning a multi-scale ontology into core and peripheral models. These models are then systematically varied to generate scenarios that may identify counter-intuitive results. This also led to the identification of a hypothesized approach to organize, navigate, and select models for reuse. However, much additional research is required and promising directions for future research are identified. Based on the results of all of the other tasks, Section 0 presents a revised and enhanced version of the enterprise modeling methodology. Finally, Section 0 concludes the report.