Robustness of decentralized decision‐making architectures in command and control systems. Issue 2 (29th October 2022)

• Record Type: Journal Article
• Authors: Boss, Lewis N.; Gralla, Erica L.
• Is Part Of: Systems engineering. Volume 26:Issue 2(2023)
• Page Start: 149
• Page End: 161
• Publication Date: 2022-10-29
• Subjects: architecture -- command -- control -- decentralization -- robustness; Systems engineering -- Periodicals; 620.0011
• DOI: 10.1002/sys.21647
• Languages: English
• ISSNs: 1098-1241

Abstract

Abstract: Organizations use command and control (C2) systems to collect, organize, and disseminate information in order to make decisions, impart instructions, and manage resources to accomplish a mission. C2 agility and robustness are critical to ensure that a C2 system can perform well in a variety of environments. One system design principle, decentralization, has been closely linked to desirable system characteristics such as agility and adaptability, but its relationship to system performance robustness is not well‐established. In this study, we explore C2 system architectures—ranging from fully centralized to fully decentralized archetypes—to assess their performance and robustness characteristics across a spectrum of operating environments. While the centralized archetype achieves high performance in favorable environmental conditions, its performance quickly degrades when conditions worsen, hindering its overall robustness. Conversely, the decentralized archetype achieves a lower but more stable performance profile resulting in more robustness when performance requirements are lower. Finally, we explore alternative, hybrid architectures with varying degrees of centralized and decentralized decision‐making. We find that by centralizing only certain system‐consequential functions such as resource allocation, and decentralizing more focused decision functions which can be performed suitably with only local information and resources, system performance and robustness are improved beyond that of the simple archetypes.