This presentation was made at the NAFEMS Americas Seminar "Model-Based Engineering: What is it & How Will It Impact Engineering Simulation" held on the 1st of October 2019 in Columbus Ohio

Resource Abstract

Behavioral modeling is common practice in the systems branch. It lies along the architecting-engineering continuum and is integral in gaining understanding of a system of interest (SoI), requirements derivation, technical optimization, and client satisfaction. Yet, the unprecedented systems that are to be pervasive in years to come will be so complex that systems modeling, alone, is not at all enough to answer stakeholder questions. Developers must consider the logical consequences of hypothetical scenarios, especially when the SoI exhibits quasi- to non- deterministic behavior. Thus, the execution of models over time (i.e. simulation) gives visibility into system design and acts as a rigorous supplement to ensure completeness & correctness. This paper/presentation discusses unique methods that can be applied when taking model-based engineering (MBE) approaches utilizing the Systems Modeling Language (SysML), the Foundational Subset for Executable UML (fUML), an appropriate development environment, and a W3C SCXML - fUML standards compliant simulation engine. Characterized by multiple levels of abstraction, these methods are defined using a broad set of constructs, formalisms, capabilities, and non-homogeneous dialects. An example case study of an integrated architecture model is discussed throughout the paper/presentation, such that the audience can better understand the application of engineering simulation within the systems branch. The investigation focuses on the development of functional and logical system architectures of platforms that exhibit ‘multi-threaded’, ‘reactive’, and ‘data-intensive’ characteristics. Examination takes place into simulation modeling of different phenomena for each of the characteristics mentioned prior. The analyses behind ‘multi-threaded’ behavior considers flexible techniques for sharing global memory resources during run-time—allowing multiple service requests to be managed with ease. The exploration of ‘reactive’ behavior outlines innovative practices to exploit comprehensive timescales—without having to rely solely on the simulation engine’s built-in clock. Lastly, the discussion on ‘data-intensive’ behavior reviews concepts of initializing large data sets at run-time, for model-based testing—with little to no effort by the architect/engineer. Approaching product development with the ideas reviewed in this paper/presentation has a high degree of utility that is desirable when attending to complex intractable questions. Cohesively defined and loosely coupled, the methods described assist in providing rapid, definitive, answers to stakeholder questions and prove pivotal when architecting-engineering dependable systems. For these types of problems, with no closed form solution, having simulation/execution contexts allows for human reasoning and mathematical analyses to complement each other—creating powerful problem-solving synergy.