Designing Resilient Systems‐of‐Systems: A Survey of Metrics, Methods, and Challenges

Abstract

Resilience is the ability of a system to react to and recover from disturbances with minimal effect on its dynamic stability. While resilience has been the focus of research in several fields, in the case of systemsof-systems (SoSs), addressing resilience is particularly interesting and challenging. As infrastructure SoSs, such as power, transportation, and communication networks, grow in complexity and interconnectivity, measuring and improving the resilience of these critical SoSs is vital in terms of safety and providing uninterrupted services. While the resilience of SoSs depends on the reliability of their constituent systems, traditional reliability and risk assessment approaches cannot adequately quantify their resilience. In this paper, we provide an evaluation of the various methods available and challenges associated with designing resilient SoSs by (1) indicating important differences between resilience and various related system attributes, (2) providing a critical assessment of the current reliability and risk techniques in addressing SoS resilience, and (3) discussing the application of recent multidisciplinary research that can guide the design of resilient SoS. Finally, we highlight key challenges in this design process and propose a series of research themes that can shape future research in this field. C⃝ 2015 Wiley Periodicals, Inc. Syst Eng 18: 491-510, 2015

Percolation‐Based Metrics to Quantify the Resilience of Communication Systems

“…The products of the design solution processes are "alternative design solutions, physical architectures, and ultimately, a final design solution" (DAU, 2018). Some representative examples of functional failure analysis include the work of Mehrpouyan (2015), Uday (2015), Kurtoglu (2007, 2008), and Jensen (2007, and Ormon (2002). Several advantages of this model are the early discovery of probable failures by associating them with product functions, use of known component failure rate to predict component failure, identification of potential functional failures, and their associated risks.…”

“…Resilience theory plays a vital role in designing engineered systems comprised of Cyber-Physical Systems (CPS) with components, sub-systems, or system interdependencies intended to make the entire system robust. Resilience is a system-level attribute closely associated with survivability, reliability, robustness, elasticity, safety, and nimbleness (Uday, 2015;Song, 2016). Although many scholars have conducted extensive studies on resilience design (Youn, 2011;Filippini, 2012;Jackson, 2012;Bhatia, 2015;and Bahun, 2016), only a few papers applied resilience to maritime systems.…”

“…Markov Analysis is a versatile tool for modeling a complex system, but the manual development of Markov Chain system behavior is a daunting task. Some representative examples of FTA, BN, and MA include the work of Henley (1981), Lewis (1994), and Boudali (2005); Uusitalo (2007), Khakzad (2013), and Uday (2015); Busic (2012) and Ericson (2016). Network Theory.…”

Sociotechnical Resilience: A Preliminary Concept

“…Further, a clear conceptualization and unique challenge of making sociotechnical systems resilient is still largely missing. Currently, the researchers who have engaged this term have adopted a systems engineering perspective with its emphasis on the technical aspects; (20,36,77,78) or, they have adopted human factors, risk, or disaster perspectives (12,37,69) with a focus on individuals and organizations. Both these perspectives, while important, produce a limited view of the challenges of sociotechnical systems.…”

Assessing causal claims about complex engineered systems with quantitative data: internal, external, and construct validity

“…These disagreements suggest that there may be multiple theoretical constructs that use the same terminology (see also, Moses, 44 who discusses the difference between computational complexity and difficulty). Similarly, there are multiple definitions of flexibility; 39,[45][46][47][48] resilience; [49][50][51] and several other "ilities." 52 It is unclear whether these concepts all index the same basic construct of changeability, 53 whether they are distinct, 46 or whether they are in some way correlated.…”