combat models, and software development. As detailed below, they are weak in the knowledge, skills, and attributes required to ensure success in this program. There is an adequate base of subject-matter expertise to design process models of sufficient detail in the areas of SLAD’s core mission, including the network operations examples described during the review. However, that level of expertise is not present in the S4 model development team, as evidenced by the lack of (1) knowledge of the current state-of-the-art in combat models; (2) a focus on temporal, spatial, and organizational scales sufficient for SLAD’s purpose; and (3) compliance with standard software development processes, including the development and documentation of requirements, the specification of an architecture and its description via flowcharts, the development of a data dictionary, and the production of users’ and programmers’ manuals.
SLAD does not appear to have personnel qualified to design and create software to support analysis of military operations, either maneuver unit warfare or wide area security. SLAD does not have personnel with the training, experience, and expertise to conduct system-of-systems analysis of military operations, to play a role in deciding what models, simulations, and tools are required. A competent staff of modeling and simulation personnel, active in the military operations research and systems analysis community, should be developed and expanded. SLAD should place a military operations researcher (OR) into a position of influence over the military mathematical modeling. Personnel development should focus on obtaining a strong operations research staff with special skills in modeling and simulation. To achieve that, it would be important to present SLAD as an exciting place to be for scientists and engineers, offering good work with a clear and important mission. It would also be important to make SLAD known and respected outside its home fields at Aberdeen and White Sands. Exhibiting SLAD’s work at conferences and other professional venues would also help to entice recent graduates to consider an ARL career.
To leverage software and modeling support, SLAD should build strong relationships with ARL and DoD modeling and simulation groups. SLAD should also expand the source and breadth of SMEs and access to more soldiers with operational experience. To establish a broader institutional system-of-systems awareness, SLAD should encourage and enable its personnel to attend, as observers, advanced warfighting experiments, simulation exercises, and command post exercises, the last of these at all echelons. In this way a foundation of experience could be built up over time and SLAD could become prepared to play its key role in collaborative system-of-systems analysis.
The current approach to S4 development is flawed. All of the S4 coders and modelers reside at the New Mexico State University Physical Sciences Laboratory (PSL); only analysts reside at SLAD. The current methodology for performing S4 operations involves SLAD analysts working with PSL to define a problem. PSL personnel then run the code and provide data to the analysts. To date, PSL has not documented its code. Furthermore, there is no program plan in place, no users’ manual, and no descriptive report with all of the details. Because of this approach, SLAD does not own the code, which is imperative for S4 success. If SLAD is unable to hire coders to assume ownership of the development and maintenance of the code, then the directorate should demand documentation and drive the direction of development through construction and enforcement of a detailed program plan.
Scope and Collaboration
To ensure that S4 will become a useful tool for the system-of-systems analyses that SLAD believes to be a requirement of SLAD’s mission space, it is important that the scope of the model be carefully chosen. The scope of engagements should be at the platoon and squad level, and certainly no larger than the company level, with interactions and dependencies on non-organic elements represented via exogenous events. It might be efficient for SLAD to consider S4 as composed of two components. The