A stakeholder framework for evaluating the ‐ilities of autonomous behaviors in complex adaptive systems. Issue 5 (19th August 2020)

• Record Type: Journal Article
• Authors: Douglas, Andre; Mazzuchi, Thomas; Sarkani, Shahram
• Is Part Of: Systems engineering. Volume 23:Issue 5(2020)
• Page Start: 633
• Page End: 655
• Publication Date: 2020-08-19
• Subjects: Systems engineering -- Periodicals; 620.0011
• DOI: 10.1002/sys.21555
• Languages: English
• ISSNs: 1098-1241

Abstract

Abstract: Complex adaptive systems (CAS) exhibit emergent behaviors caused by the nonlinear actions between individual components within the system and their environment. These emergent behaviors are difficult to predict and measure making it very challenging to initially design from the start. Developing CASs, such as an autonomous swarming unmanned system, is a relevant key task today for solving hard problems with limited resources in a demanding economy. Swarming unmanned systems are valued by their unique capabilities, but how do stakeholders ensure their needs are considered during development? Certain stakeholders have interests that go beyond the initial delivery of the system, yet developers seem to focus efforts on designing autonomous behaviors to satisfy immediate mission needs. To close this gap, we propose a stakeholder focused framework for evaluating CASs over a full system life cycle for swarming unmanned systems. A goal‐oriented requirements engineering modeling methodology, Knowledge Acquisition in AutOmated Specification, is used to map requirements between stakeholder goals and autonomous behaviors. Behaviors are analyzed in an agent‐based simulation using NetLogo for quantifying stakeholder‐driven performance qualities (‐ilities). A case study was performed for a swarm of unmanned surface vessels called wave gliders to evaluate performance over the entire system life cycle.