A Markov Perspective on Development Interdependencies in Networks of Systems

Abstract

The development and acquisition of complex systems remain a challenge, especially in the aerospace/defense sector, due to complexities in both program management and engineering design that often are a result of interdependencies. The interdependencies between constituent systems form networks that, while enabling capabilities that are beyond those of individual systems, also increase risk since disruptions in the development of one system may propagate to other directly or indirectly dependent systems. This paper demonstrates an approach to aggregate the network interdependency characteristics and compare alternatives with respect to the time required to arrest the propagation of development delays in a network. Delay propagation is modeled as a Markov chain, where the states are defined as the constituent systems and the transition probabilities as the dependency strengths between systems. A proof-of-concept application shows the approach can distinguish between alternate networks and indicates its applicability for managing risk in design and development of systems with significant interdependencies.

A computational concept generation method for a modular vehicle fleet design

“…Finally, we intend to further extend this work by characterizing the links made between functions to create modules, allowing for an evaluation of module or vehicle complexity and flexibility as in [24] or [25]. This characterization will allow for evaluation of fleet architectures where certain links are promoted or inhibited during the synthesis process as described here, improving understanding of how module design choices impact fleet performance.…”

Assessing the Impact of Development Disruptions and Dependencies in Analysis of Alternatives of System-of-Systems

“…The metric is developed according to the formulation by (Mane, DeLaurentis, & Frazho, 2011). A typical network is shown in Figure 27.…”

“…. The expected time before disruption absorption is computed for each node, by Markov property (Mane, DeLaurentis, & Frazho, 2011), as…”

“…Works in by (Mane M. C., 2012) on the other hand, adopt an operations research (OR) perspective through concurrent engineering paradigms that seek to improve efficiency by reducing decision 'handoffs' between interdependent systems; namely between airline operations and aircraft design. Research by (Mane, DeLaurentis, & Frazho, 2011) pursues a network driven Markov perspective to SoS complexity and attempts to quantify cascading effects of single nodal disruptions on a network. Literature by (Garrett, 2011)has presented works on managing interstitial spaces for a SoS case of the Ballistic Missile Defense System [BMDS]; the work focuses on issues of interoperability, integration and domain interfaces through employment of graph theoretic and agent based methods, coupled with developed managerial and technical metrics.…”

Supporting design via the System Operational Dependency Analysis methodology

“…Moreover, the behavior of most systems designed and developed nowadays cannot always be directly evaluated based only on the behavior of the individual systems or subsystems. In this paper, we define complex systems as entities where interactions between the parts have a fundamental impact on the global behavior, and the effect of dependencies is difficult to ascertain by simple inspection during design (Mane et al 2011;Nai Fovino and Masera 2006;Hsu et al 2009). Throughout this paper, we will use the term systems to refer to this class of complex systems.…”