Systems of Systems Analytic Approaches
Lead Authors: Mike Henshaw, Judith Dahmann
A review of SoS IEEE publications on SoS from 2021-2023 [Dahmann, 2024] looked across the SoS papers published in IEEE from 2020-2023 addressing a range of domains, incorporated several technical approaches including model-based approaches, graph analysis and artificial intelligence.
Model-based Approaches
First, as systems engineering moves towards becoming a model-based discipline, SoS engineering is part of this movement. Looking across the 138 domain focused papers in this review, the vast majority (120 or 90.9%) apply some type of model-based approach. These model-based approaches take different forms, including:
Network Models
These models represent the system of systems (SoS) as a network of interconnected nodes. They are used to analyze the robustness, resilience, and other properties of the system. Examples include the combat network model, network modeling of kill-web, and network model of a weapon technology system-of-systems.
Agent-Based Models
These models represent the SoS as a collection of agents, each with its own behavior. They are used to simulate the behavior of the system and evaluate policy changes. Examples include the agent-based modeling to represent the kidney transplant system of systems [Threlkeld, R et al. 2022].
Mathematical and Optimization Models
These models represent the SoS using mathematical equations or optimization problems. They are used to analyze and optimize the system's performance. Examples include the mixed-integer programming model for the combat system-of-systems architecture design problem and the mathematical model for the set covering problem.
Model-Based Systems Engineering (MBSE) and SysML
These tools are used to describe the structure and behavior of the SoS. They support activities like requirement analysis, design, verification, operation, and maintenance activities.
Simulation Models
These models are used to simulate the behavior of the SoS under different conditions. Examples include the simulation models to represent the kidney transplant system of systems [Threlkeld, R et al. 2022] and the digital real-time simulators.
Machine Learning Models
These models use machine learning algorithms to predict or classify outcomes based on input data. Examples include the use of machine learning algorithms to predict building maintenance and optimize energy use, and the use of a supervised learning model for classification and prediction of the failure status system.
Conceptual Models
These models represent the high-level structure and concepts of the SoS. Examples include the conceptual model for an interoperable and vendor-neutral communication framework for the agricultural domain and the conceptual model based on Hierarchical Systems (HS) technology.
Graph Analytic Approaches
By their nature networks of interconnected systems, SoS lend themselves to graph analytic approaches. From this set of SoS domain papers, 29 (of 132 or 22%) of the domain application papers employ graph analysis in some form. This includes:
Network Modeling
This includes modeling SoS architectures as single-layer or multilayer networks, using complex network methods for robustness assessment, and using network modeling and adjacency matrices to represent the kill-web.
Hypergraph Theory
This includes using hypergraphs to model the system of systems and represent the constraints between subsystems and utilizing hypergraph theory to manage the evolutionary development property in the complicated internal structure of SoS during physical faults.
Graph Theory
This includes applying graph theory and social network analysis to model communications, healthcare and other types of systems-of-systems.
Knowledge and Information Representation
This includes using graphical representations for the organizational modeling of the system of systems.
Dependency and Interconnection Analysis
This includes using Design Structure Matrices (DSM) for analyzing interdependencies of elements within a SoS.
Complex Network Analysis
This includes using complex network theory to Construction of a combat network based on the kill web, which can be represented as a graph with nodes and edges.
Artificial Intelligence Approaches
Finally, a quarter of these papers (36 out of 132, or 26%) applied artificial intelligence (AI) in their SoS analysis approach. The AI approaches used in these papers can be categorized as follows:
Machine Learning Algorithms
These include general machine learning techniques for fault detection and isolation, feature selection, and data processing. An example is the paper: "Federated Feature Selection for Cyber-Physical Systems of Systems" [P. Cassará et al. 2022].
Neural Networks and Deep Learning
This category includes applications to human-machine collaboration and swarm techniques for search and rescue operations with example papers: "A Robotic System of Systems for Human-Robot Collaboration in Search and Rescue Operations" [Chan T. H. et al. 2023] and "A Capability Fitting and Data Reconstruction Model Based on Particle Swarm Optimization-Bidirectional Deep Neural Network for Search and Rescue System of Systems" [Gao, Y. et al. 2023].
Genetic Algorithms
Genetic algorithms are used for mission planning, sensor allocation, and meta-architecture optimization. An example here is "A System of Systems for the Optimal Allocation of Pollutant Monitoring Sensors" [C. Carnevale. C. et al. 2022]
Artificial Intelligent Agents
The concept of Non-Human Knowledge Workers is introduced, which are artificial intelligent agents designed to relieve human knowledge workers of cognitive tasks in the one paper in this category: "A New Technology Rises Non-Human Knowledge Workers and Decision-Making in a System of Complex Systems" [Mortimore D. Et al, 2023]
Fuzzy Logic and Inference Systems
Fuzzy Inference Systems are used to assess the overall fitness measure of the System of Systems (SoS) and to calculate the overall fitness value in a genetic algorithm. An example here is "System of Systems Meta-Architecture Approach to Improve Legacy Metrorails for Enhanced Customer Experience" [Polley, M. et al. 2021].
AI in Autonomous Systems
The papers discuss the use of AI and machine learning in enhancing the level of autonomy in systems, including autonomous vehicles and offshore wind farms, as presented in "Symbiotic System of Systems Design for Safe and Resilient Autonomous Robotics in Offshore Wind Farms" [Mitchell, D. et al. 2021].
References
Works Cited
Threlkeld, R., Ashiku, L., & Dagli, C. (2022). Complex System Methodology for Meta Architecture Optimization of the Kidney Transplant System of Systems. In 2022 17th Annual System of Systems Engineering Conference (SOSE). IEEE. DOI: 10.1109/SOSE55472.2022.9812668
Cassará, P. A. Gotta, and L. Valerio, "Federated Feature Selection for Cyber-Physical Systems of Systems," in IEEE Transactions on Vehicular Technology, vol. 71, no. 9, pp. 9937-9950, September 2022, doi: 10.1109/TVT.2022.3178612.
Chan, T. H., Halim, J. K. D., Tan, K. W., Tang, E., Ang, W. J., Tan, J. Y., Cheong, S., Ho, H.-N., Kuan, B., Shalihan, M., Liu, R., Soh, G. S., Yuen, C., Tan, U.-X., Heng, L., & Foong, S. (2023). A Robotic System of Systems for Human-Robot Collaboration in Search and Rescue Operations. In 2023 IEEE/ASME International Conference on Advanced Intelligent Mechatronics (AIM) (pp. 878-885). IEEE. DOI: 10.1109/AIM46323.2023.10196185
Gao, Y., Liu, H., Niu, F., & Tian, Y. (2023). A Capability Fitting and Data Reconstruction Model Based on Particle Swarm Optimization-Bidirectional Deep Neural Network for Search and Rescue System of Systems. IEEE Access, 11, 10366-10383. doi:10.1109/ACCESS.2023.100083
C. Carnevale, L. Sangiorgi, E. De Angelis, R. Mansini, and M. Volta, "A System of Systems for the Optimal Allocation of Pollutant Monitoring Sensors," in IEEE Systems Journal, vol. 16, no. 4, pp. 6393-6400, December 2022.
Dahmann, Judith (2023). “Current Landscape of System of Systems Engineering”, IEEE Systems of Systems Conference; Tacoma, Washington.
Mortimore, D., Aten, K., & Buettner, R. R. (2023). A New Technology Rises: Non-Human Knowledge Workers and Decision-Making in a System of Complex Systems. In Proceedings of the 2023 18th Annual System of Systems Engineering Conference (SoSe) (pp. 1-10). IEEE. DOI: 10.1109/SOSE59841.2023.10178624
Experience. In 2021 16th International System of Systems Engineering Conference (SoSE) (pp. 191-196). IEEE. DOI: 10.1109/SOSE52739.2021.9497492
Mitchell, D., Blanche, J., Zaki, O., Roe, J., Kong, L., Harper, S., Robu, V., Lim, T., & Flynn, D. (2021). Symbiotic System of Systems Design for Safe and Resilient Autonomous Robotics in Offshore Wind Farms. IEEE Access, 9, 141421-141445. https://doi.org/10.1109/ACCESS.2021.3095331