IDENTIFICATION OF BEHAVIOR PATTERNS IN SYSTEMS-OF-SYSTEMS ARCHITECTURES

This dissertation presents a methodology to derive inherent behavior patterns in system-of-systems architectures. When considering the possibility of not only the intended behaviors of a system but also the alternative, unintended behaviors of the system, the developer needs to evaluate hundreds of possible outcomes, even for a simple model. This set of outcomes grows to tens of thousands for a reasonably complex model, driving the need for some means of systematic analysis that ensures the evaluation of all possibilities. This dissertation outlines a methodology that enables the developer to describe the behaviors of a particular system; define the interaction of the system with other systems, users, and the environment; add constraints to the model; derive patterns from the set of possible outcomes; analyze and interpret the results; and re-use the model for other problems of interest. The methodology employs the Monterey Phoenix behavior modeling concepts and tools as a light-weight, formal method. The dissertation details a cross-domain example that demonstrates the methodology and outlines multiple applications, serving as verification of the approach.