Computational Intelligence Approach To SoS Architecting And Analysis

Abstract

"System-of-Systems" (SoS) is an emerging and essential multidisciplinary area of systems engineering impacting various systems across multiple disciplines. SoS are difficult to optimize and understand due to their adaptive emergent behaviors, dynamically changing system boundaries, nonlinear inputs and outputs, as well as multiple stakeholders with competing goals. The individual systems alone cannot independently achieve the overall goal of the SoS and are dependent upon each other. The constant interaction between the systems and the interdependencies produces emergent properties that cannot be fully accounted for, analyzed, and optimized by the "normal" systems engineering practices and tools. SoS engineering (SoSE) is an emerging discipline in systems engineering that attempts to create methodologies for approaching SoS problems in various disciplines. The SoS architecture organization of these systems involves many of web-like connections and demonstrates the ability of individualized adaptability. Meta-architecting can help achieve the optimized architecture for these complex SoS. A form of meta-architecting is the Flexible and Intelligent Learning Architectures for System of Systems (FILA-SOS) utilizes a straightforward system definitions methodology and an analysis framework that supports the exploration and understanding of the key trade-offs and requirements by a broad range of SoS stakeholders and decision-makers. FILA-SoS and the Wave Process address SoS architecting's most challenging aspects and pain points. Developing models of acknowledged SoS architectures can assist in discovering and defining issues, satisfying problems with stakeholder needs, and analyzing the impact of policies through the rules on architecture selection. Key performance attributes that depend on architecture selection can be discovered through facilitated interactions with stakeholders and subject-matter experts (SMEs). Relatively simple fuzzy rule-based systems can be created and combined with the key performance attributes (KPAs) evaluations for an overall SoS assessment. A fuzzy genetic approach is utilized for finding solutions to several SoS architecting problems integrated with a restrictive meta-model of undirected network graphs representing the system interfaces. Defining the boundaries of the membership functions and changing them independently is an excellent way to quickly get answers about the problem. Following the FILA-SoS makes it relatively easy to switch back to the real values to perform a "what-if analysis." Two use cases highlighted in this chapter are cybersecurity and healthcare, but a list of many more use cases in research are given at the end of the chapter. Both use cases utilized the SoS Explorer. For the health care use case, a meta-architecture was created for the organ procurement SoS to optimize for selecting the best participating systems for a given set of donor kidneys. Utilizing the KPAs, fuzzy membership functions, rules, and genetic algorithms, two use cases were optimized using a fuzzy inference to find the best meta-architecture for the healthcare SoS. The rules of the fuzzy inference engine represent non-linear trade-offs associated with KPAs of multiple stakeholders. For the cybersecurity use case, meta-architecting and optimization is demonstrated through connecting genetic algorithm with fuzzy inference system (FIS) to provide an intuitive and interactive visualization of meta-architectures.

Department(s)

Engineering Management and Systems Engineering

Keywords and Phrases

Architecting; Artificial intelligence; Meta-architecture; Optimization; System of systems

International Standard Book Number (ISBN)

978-139420331-4;978-139420328-4

Document Type

Article - Journal

Document Version

Citation

File Type

text

Language(s)

English

Rights

© 2023 Wiley, All rights reserved.

Publication Date

29 Sep 2023

Link to Full Text

Share

 
COinS