Experts share insights during Executive Plenary Panel during the 2018 IEEE International Systems Conference in Vancouver, Canada.
John Blyler, Founder, JB Systems Tech and Media
The changing shape of system complexity was the focus of the Executive Plenary Panel at this year’s IEEE SysCon 2018 event in Vancouver, BC. Prior to beginning the panel, the moderator and technical chair for the conference, Robert Rassa from the Raytheon Company, provided several examples of how customers influenced the systems engineering process.
“The more complex the system, the more difficult its development becomes,” observed Rassa. “The customer defines the way you’ll go about your system engineering effort. Being able to deal with this challenge is an important aspect of systems engineering.”
Following Rassa’s introduction to the topic, each panelist provided insights concerning what adds complexity to today’s systems.
The first panelist was Dr. Sidney Givigi from the Royal Military College of Canada. He observed that, in the past, most systems were developed in a model-based manner by combining different models, for example, a model of a vehicle combined with a model from sensors. A significant challenge was the manner in which different models would interact with each other. Compounding the complexity was the addition of the data component of the system.
“The models and the data driven systems must be interconnected,” explained Givigi. “The traditional way of accomplishing this was to treat the data driven part as a black box.” But this approach didn’t work in applications where the designer had to guarantee safety boundaries and/or operational limits, as in autonomous vehicles, for instance. For these situations, machine learning and other adaptive approaches must be combined with the data driven systems.
Givigi emphasized that systems engineering is required to determine the most effective way to put the components together, for example, black box data combined with rule-based systems, machine learning, and even artificial intelligence.
The next panelist added the human element to the discussion.
“In recent years, I’ve been looking more into the social-technical issues that surround the discipline of systems engineering,” explained Dr. Donna Rhodes from the Massachusetts Institute of Technology (MIT) and a past-president of INCOSE. “There are many technical issues, but we have some daunting issues that relate more to people.”
Her research indicated that social factors often dominate over technical factors when making decisions using models. For example, there are contentious situations in many organizations between very experienced senior people and far less experienced but model and technology savvy younger people. While the novices don’t particularly understand the foundations of the corporate decision making process, the senior members don’t fully understand the tools or technologies. Rhodes stressed the need to overcome these kind of culture issues in order to fully make good use of model-centric technologies.
The next panelist focused on the complexities added by cybersecurity issues. Speaking in his capacity as a private citizen (and not as a Lieutenant Colonel in the US Air Force), Logan Mailloux worried that today’s mission critical systems may not function as intended when subjected to a highly contested cybersecurity environment. His primary concern centered on globally developed and manufactured COTS technologies, uncertainties in the supply chain as to the origin of systems, and system inter-connections that obfuscate possible system states and vulnerabilities.
“Typically, we deal with complexity via standardization,” explained Mailloux. He identified three important DOD documents that help to address system-level, cybersecurity issues:
- Department of Defense Instruction (DoDI) 5000.02 – Operation of the Defense Acquisition System, Enclosure 14: Cybersecurity in the Defense Acquisition System.
- Defense Acquisition Guidebook (DAG); Chapter 3, Systems Engineering and Chapter 9, Program Protection Planning.
- Department of Defense (DOD) Risk, Issue, Opportunity (RIO) Management Guide for Defense Acquisition Programs.
The next panelist focused on emergent systems.
“Originally, systems engineering didn’t include the word “complex,” noted Dr. Paul Hershey of Raytheon.” It’s important to remember that today’s complex systems engineering activities include complexity, systems, and engineering.”
He emphasized that complexity involves interconnected and interwoven components that can be defined in terms of emergence, a new behavior that develops as a result of the interactions among the component systems. Emergent behavior cannot be deduced from the behaviors of the individual systems themselves, whether considered individually or in subgroups.
Several examples of emergent behavior were provided in the area of big data analytics, where new forms of processing are needed to enable enhanced decision-making.
The last panelist, Dr. Vincenzo Piuri from the University of Milan, Italy, and technical conference chair, added that today’s complex systems often need to make decisions autonomously and resiliently, for example, to dynamically adjust a system to deal with degradation while aiming for maintenance. He emphasized the need for university education and continuing engineering education to deal with the complexity of today’s systems.