An Introduction to Model-Based Systems Engineering and Hetero-functional Graph Theory for Hydrological Systems of Systems
Listen
Radio episode about this paper
Transcript
Introduction to the show: ident: Robotics Radio. Generated commentary on the latest robotics and control papers.
Rosa: We're looking at the paper: "An Introduction to Model-Based Systems Engineering and Hetero-functional Graph Theory for Hydrological Systems of Systems".
Dev: Here's the gist of the paper. An introduction to model-based systems engineering and hetero-functional graph theory for hydrological systems provides a unified modeling framework that explicitly represents system structure and function through an ontology, making complex environmental system analysis computationally tractable Core Methodology The paper introduces a unified modeling framework for representing interconnected environmental systems using Model-Based Systems Engineering and Hetero-functional Graph Theory.
Rosa: This methodology formalizes knowledge from existing domain-specific process models into a consistent ontology that explicitly represents system structure, function, and interconnections.
Dev: The approach combines Model-Based Systems Engineering (MBSE) with HFGT to formally represent system elements, processes, and their interactions within a consistent modeling framework.
Rosa: Conceptual Elements of HFGT The HFGT meta-architecture represents every engineering system using five fundamental concepts: operands, resources, processes, buffers, and capabilities.
Paper summary: Rosa: We just talked about how this paper, "An Introduction to Model-Based Systems Engineering and Hetero-functional Graph Theory for Hydrological Systems of Systems", is trying to formalize knowledge from existing environmental models. The thesis is that you can take the structural and process level information that’s often hidden in those specific domain formulations and turn it into a consistent ontology.
Dev: And this ontology explicitly represents the system structure, function, and how everything connects. It claims they use Model Based Systems Engineering combined with Hetero functional Graph Theory to do this unification.
Taro: So it’s not just describing what the system does, but mapping out the entire relationship map between the components and their activities in a very formal way.
Rosa: Exactly. They show how conventional process models can be translated into a scalable system of systems representation using these hydrological examples, which helps make sense of how interconnected those things really are across different scales.
Dev: The paper shows that this methodology establishes a systematic workflow for watershed modeling, and it demonstrates translation from the system architecture into mathematical models. It illustrates how this works with progressively complex watershed examples.
Taro: I’m interested in the part where they show how they establish best practices for ontology driven environmental systems modeling, because that suggests there are established rules for doing this kind of mapping properly.
Rosa: That’s right, and it sets up a system of systems representation that is distinct from single domains under a common ontology, which is pretty neat.
Dev: The main point they are making here is how to move from implicit knowledge in those process models to explicit knowledge in the structure itself, using this combined approach.
Taro: It seems like they are providing a way to bridge the gap between high level architectural design and the detailed mathematical simulations that follow.
Conclusion: Rosa: So looking at "An Introduction to Model-Based Systems Engineering and Hetero-functional Graph Theory for Hydrological Systems of Systems" overall, what we’ve seen is a systematic workflow that translates conceptual descriptions into executable mathematical models by making the system structure explicit.
Dev: It provides this structured framework for integrating traditional process based modeling with explicit structural representation, which is particularly useful because you can validate things incrementally before introducing additional resources or processes.
Taro: The implication is that you get a repeatable workflow that can be adapted to many other application domains beyond just watershed applications, because the core methodology isn't tied to one specific type of system.
Rosa: Exactly. It’s not meant to replace traditional engineering models, but rather provide this structured framework for integrating them by making the system form and function explicit so you can see how things fit together.
Dev: So ultimately, it’s about synthesizing increasingly complex environmental systems while preserving structural transparency in the process of modeling them.
Megan S. Harris, Mohammad Mahdi Naderi, Ehsanoddin Ghorbanichemazkati, John C. Little, Amro M. Farid
Virginia Tech Department of Civil and Environmental Engineering · Stevens Institute of Technology
eess.SY, cs.SY
Submitted: 2025-05-31
Updated: 2026-11-03
Code: https://github.com/LIINES/HFGT-Hydrology-Model
License: http://creativecommons.org/licenses/by/4.0/
Importance score: 77/100
The gist: An introduction to model-based systems engineering and hetero-functional graph theory for hydrological systems provides a unified modeling framework that explicitly represents system structure and…
Key concepts
- Model-Based Systems Engineering (MBSE)
- MBSE is a process that focuses on building digital models of engineering systems first. It involves creating a consistent reference architecture to capture the system's structure and capabilities before applying complex mathematical physics, ensuring the model accurately reflects how components interact.
- Hetero-functional Graph Theory (HFGT)
- HFGT is a meta-architecture that uses five core concepts—operands, resources, processes, buffers, and capabilities—to define engineering systems. It formally represents how assets are operated upon by specific actions within the system structure.
- Hetero-functional Incidence Tensors (HFITs)
- These tensors are mathematical tools used to capture the heterogeneity of engineering systems. They explicitly map which operands are acted upon by which capabilities at specific buffer locations, providing a rigorous way to represent system flows and interactions mathematically.
- Engineering System Net (ESN)
- The ESN is a Petri-net representation that tracks the state of an engineering system over time. Its state transition function acts like a generalized conservation equation, directly modeling mass balance in hydrological systems by tracking operand additions and removals.
Terminology
Summary
An introduction to model-based systems engineering and hetero-functional graph theory for hydrological systems provides a unified modeling framework that explicitly represents system structure and function through an ontology, making complex environmental system analysis computationally tractable
Core Methodology
The paper introduces a unified modeling framework for representing interconnected environmental systems using Model-Based Systems Engineering and Hetero-functional Graph Theory. This methodology formalizes knowledge from existing domain-specific process models into a consistent ontology that explicitly represents system structure, function, and interconnections. p. 2 Existing environmental models encode rich structural and
process-level information, yet much of this knowledge remains implicit within domain-specific formulations. This paper demonstrates how such information can be formalized into a consistent ontology that explicitly represents system structure, function, and
interconnections. Rather than replacing existing process models, the methodology provides a
common representational framework that makes their underlying structure explicit while preserving the scientific knowledge
embedded within them. The approach combines Model-Based
Systems Engineering (MBSE) with HFGT to formally represent system elements, processes, and their interactions within
a consistent modeling framework [12, 18].Where the paper says it
Where the paper says it
Conceptual Elements of HFGT
The HFGT meta-architecture represents every engineering system using five fundamental concepts: operands, resources, processes, buffers, and capabilities. System Operands are defined as assets or objects that are operated on or consumed during the execution of a process. p. 5 HFGT refers to these entities as operands li ∈ L. Definition 8 – System Operand [30]: An asset or object li ∈
L that is operated on or consumed during the execution of a
process. System Processes
Operands are transformed and transported throughout the
system in what HFGT refers as processes pw ∈ P. Definition 9 – System Process [31, 30]: An activity p ∈ P that
transforms or transports a predefined set of input operands into
a predefined set of outputs. Resources are the physical or logical
entities that perform processes within the system. Definition 10 – System Resource [30]: An asset or object rv ∈
R that facilitates the execution of a process. Capabilities
Different resources can perform different sets of processes. HFGT represents these allowable actions as capabilities.Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Mathematical Formulation
The conceptual ontology is translated into a mathematical representation through hetero-functional incidence tensors, Engineering System Nets, and the Hetero-functional Network Minimum Cost Flow formulation. Hetero-functional incidence tensors (HFITs) capture the heterogeneity of engineering systems by explicitly representing which operands are acted upon by which capabilities at which buffer locations. p. 7 Unlike traditional graph incidence matrices, which represent
connectivity between homogeneous nodes and edges, HFITs
capture the heterogeneity of engineering systems by explicitly
representing which operands are acted upon by which capabilities at which buffer locations. Definition 13 – The Negative 3rd Order Hetero-functional
Incidence Tensor (HFIT Mf−
ρ
[22]): The negative heterofunctional incidence tensor Mgρ
−
∈ {0, 1}
|L|×|BS |×|ES |
is a thirdorder tensor whose element Mf−
ρ
(i, y, ψ) = 1 when the system capability ϵψ ∈ ES pulls operand li ∈ L from buffer
bsy ∈ BS . The net incidence matrix is then
defined as:
M = M+ − M−
. (3)
This transformation provides the mathematical representation of operand flows through the engineering system.Where the paper says it
Where the paper says it
System Dynamics and Simulation
The structural relationships are represented in an Engineering System Net (ESN), a Petri-net representation that tracks operand storage and capability execution through discrete time. The State Transition Function of the ESN is equivalent to a generalized conservation equation, where the state of both the set of places and the set of transitions evolves according to the difference between operand additions and removals caused by executing system capabilities. p. 7 ■
The Engineering System Net State Transition Function is
equivalent to a generalized conservation equation. Specifically,
the state of both the set of places and the set of transitions
evolves according to the difference between operand additions
and removals caused by executing system capabilities. QB[k + 1] = QB[k] + M+U
+
[k]∆T − M−U
−
[k]∆T (6)
QE[k + 1] = QE[k] − U
+
[k]∆T + U
−
[k]∆T (7)
where ∆T is the duration of the simulation time step. ■
For hydrological systems, Equation 6 corresponds directly
to a mass balance, where buffer storage values in QB represent
water volumes or constituent nitrogen masses, and capability
firings in U represent flows of water or nitrogen between storage locations.Where the paper says it
Where the paper says it
Scalability and Application
The methodology demonstrates scalability through progressively complex watershed examples, starting from a single-lake configuration and extending to three lakes with land segments. The separation of system structure from system physics allows the same structural representation to support different physical models without modifying the underlying network formulation. p. 8 Hetero-functional Network Minimum Cost Flow Formulation
The Engineering System Net provides a dynamic representation of system evolution through operand storage and capability execution. The Hetero-functional Network Minimum Cost
Flow (HFNMCF) formulation extends this representation into
a constrained optimization framework that determines feasible
system behavior while satisfying conservation laws, capability
relationships, boundary conditions, and physical device models. Although the examples focus on hydrological systems, the methodology is equally
applicable to a broad range of environmental, engineered, and
socio-technical systems. Throughout the tutorial, references to
applications in domains such as energy systems, water infrastructure, transportation, manufacturing, and healthcare are provided to illustrate the generality of the approach and guide readers toward more advanced applications.Where the paper says it
Where the paper says it
Key Takeaways for Modeling
The key takeaways emphasize modeling form and function before mathematics, establishing the reference architecture first to capture structural elements and capabilities. The separation of conservation laws from constitutive behavior is a principal strength of the HFNMCF formulation, enabling the use of discipline-specific process models without altering the underlying mathematical framework. p. 8 The Hetero-functional Network Minimum Cost
Flow (HFNMCF) formulation extends this representation into
a constrained optimization framework that determines feasible
system behavior while satisfying conservation laws, capability
relationships, boundary conditions, and physical device models. This section defines the formulation of the HFNMCF problem while deriving the domain-specific formulation for the described hydrological systems. One barrier to developing such integrated models is the lack
of a shared conceptual language for describing systems. Models from different disciplines often define similar concepts using
different terminology or represent comparable processes using
incompatible abstractions. While using simplified examples, the overview provides a foundation for
integrating heterogeneous models across environmental and engineered domains. More broadly, this work offers a step-by-step
introduction to a generalizable modeling framework for complex systems.Where the paper says it
Where the paper says it
Where the paper says it
Conclusion
The paper presents a systematic workflow that translates conceptual system descriptions into executable mathematical models, making the underlying system structure explicit through the MBSE–HFGT methodology. This approach provides a structured framework for integrating traditional process-based modeling with explicit structural representation, suitable for many other application domains. p. 3 Rather than replacing existing process models, the methodology provides a
common representational framework that makes their underlying structure explicit while preserving the scientific knowledge
embedded within them. The approach combines Model-Based
Systems Engineering (MBSE) with HFGT to formally represent system elements, processes, and their interactions within
a consistent modeling framework [12, 18]. Little, Amro M. Farid
• Introduces a systems engineering workflow for watershed
modeling
• Defines hetero-functional graph theory with watershed examples
• Demonstrates translation from system architecture to
mathematical models
• Illustrates scalable modeling through progressively complex watershed examples
• Establishes best practices for ontology-driven environmental systems modeling Rather than replacing existing process models, the methodology provides a
common representational framework that makes their underlying structure explicit while preserving the scientific knowledge
embedded within them. The approach combines Model-Based
Systems Engineering (MBSE) with HFGT to formally represent system elements, processes, and their interactions within
a consistent modeling framework [12, 18]. Understanding these interactions therefore requires modeling approaches capable of representing not only individual systems,
but also the larger systems of systems in which they are embedded. As these systems become increasingly interconnected,
there is a growing need for methodologies that can consistently represent multiple interacting systems while preserving
the unique characteristics of each domain [1, 3, 4, 5].Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Acknowledgments
This research is based on work supported by the Growing Convergence Research Program of the National Science Foundation under Grant Numbers OIA 2317874 and OIA 2317877. All software and data used to generate the illustrative simulation results are openly available via a public GitHub repository at https://github.com/LIINES/HFGT-Hydrology-Model. p. 4 As a result, the HFGT concepts introduced using this pedagogical example generalize naturally to larger watersheds and, more broadly, to other classes of engineering systems. The objective of this tutorial is not to replace the governing equations of hydrology, but rather to demonstrate how an
equivalent mathematical model can be derived systematically
from an explicit representation of system structure and function. As a result, the HFGT concepts introduced using this pedagogical example generalize naturally to larger watersheds and, more broadly, to other classes of engineering systems. The objective of this tutorial is not to replace the governing equations of hydrology, but rather to demonstrate how an
equivalent mathematical model can be derived systematically
from an explicit representation of system structure and function. ■
Decades of progress in environmental modeling have produced sophisticated process models for representing hydrological, geochemical, ecological, atmospheric, and other Earth
systems [8, 9]. These models provide detailed, theory-driven
representations of individual systems and encode knowledge
of system structure, process connectivity, and causal relationships beyond the mathematical equations used to simulate
them [8, 9, 10, 11]. ■
Decades of progress in environmental modeling have produced sophisticated process models for representing hydrological, geochemical, ecological, atmospheric, and other Earth
systems [8, 9]. These models provide detailed, theory-driven
representations of individual systems and encode knowledge
of system structure, process connectivity, and causal relationships beyond the mathematical equations used to simulate
them [8, 9, 10, 11]. 4 represent the column indices. Therefore, the net incidence matrix, M, for the hydrological
system takes a block matrix form:
M =
M+
1,1
0 0 M1,4 0 0 M+
1,7
0 M1,9 0
0 M+
2,2
0 0 M2,5 0 M−
2,7
0 0 0
0 0 0 0 0 M3,6 0 0 M3,9 0
0 0 0 M4,4 0 0 0 M+
4,8
0 M4,10
0 0 M+
5,3
0 M5,5 0 0 M−
5,8
0 0
0 0 0 0 0 M6,6 0 0 0 M6,10
(4)
The convention MB,ψ is adopted to reflect an incidence matrix
block corresponding to the operand-buffer pair row index B and
the ψ capability index. 4 represent the column indices. Therefore, the net incidence matrix, M, for the hydrological
system takes a block matrix form:
M =
M+
1,1
0 0 M1,4 0 0 M+
1,7
0 M1,9 0
0 M+
2,2
0 0 M2,5 0 M−
2,7
0 0 0
0 0 0 0 0 M3,6 0 0 M3,9 0
0 0 0 M4,4 0 0 0 M+
4,8
0 M4,10
0 0 M+
5,3
0 M5,5 0 0 M−
5,8
0 0
0 0 0 0 0 M6,6 0 0 0 M6,10
(4)
The convention MB,ψ is adopted to reflect an incidence matrix
block corresponding to the operand-buffer pair row index B and
the ψ capability index. ■
Decades of progress in environmental modeling have produced sophisticated process models for representing hydrological, geochemical, ecological, atmospheric, and other Earth
systems [8, 9]. These models provide detailed, theory-driven
representations of individual systems and encode knowledge
of system structure, process connectivity, and causal relationships beyond the mathematical equations used to simulate
them [8, 9, 10, 11]. ■
Decades of progress in environmental modeling have produced sophisticated process models for representing hydrological, geochemical, ecological, atmospheric, and other Earth
systems [8, 9]. These models provide detailed, theory-driven
representations of individual systems and encode knowledge
of system structure, process connectivity, and causal relationships beyond the mathematical equations used to simulate
them [8, 9, 10, 11]. While using simplified examples, the overview provides a foundation for
integrating heterogeneous models across environmental and engineered domains. More broadly, this work offers a step-by-step
introduction to a generalizable modeling framework for complex systems. Throughout the tutorial, references to
applications in domains such as energy systems, water infrastructure, transportation, manufacturing, and healthcare are provided to illustrate the generality of the approach and guide readers toward more advanced applications. Paper Outline
The remainder of this tutorial is organized as follows.Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
References
The references section lists all external works cited throughout the paper, including foundational texts and recent research on system-of-systems engineering and environmental modeling. The list includes work on political impact of sustainable development goals, complex adaptive human-earth systems, and specific applications in water quality modeling. p. 2 While using simplified examples, the overview provides a foundation for
integrating heterogeneous models across environmental and engineered domains. More broadly, this work offers a step-by-step
introduction to a generalizable modeling framework for complex systems. Keywords: environmental modeling, systems modeling, modeling framework, Model-Based Systems Engineering (MBSE),
Hetero-functional Graph Theory (HFGT), system of systems
Learning Objectives
After reading this tutorial, the reader will be able to:
• Explain the need for unified modeling frameworks in
system-of-systems problems. • Describe the step-by-step methodology for transforming a physical system into a Hetero-functional Graph
Theory (HFGT) model suitable for analysis and simulation. Once represented
within this common ontology, heterogeneous system models
can be analyzed using a unified mathematical framework while
remaining faithful to their domain-specific formulations. Definition 3 – Hetero-functional Graph Theory (HFGT [22,
23]): A modeling methodology that bridges Model-Based Systems Engineering (MBSE) and network science by preserving
the heterogeneous conceptual and ontological constructs of engineering systems within graph-based mathematical representations for analysis and simulation ■
The objective of this tutorial is to introduce the methodology rather than provide a comprehensive treatment of HFGT. Paper Objective
This tutorial introduces Hetero-functional Graph Theory
(HFGT) as a methodology for constructing ontology-based
models of environmental systems of systems. Rather than replacing existing process models, the methodology provides a
common representational framework that makes their underlying structure explicit while preserving the scientific knowledge
embedded within them.Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Software and Data Availability
All software and data used to generate the illustrative simulation results are openly available via a public GitHub repository at https://github.com/LIINES/HFGT-Hydrology-Model. The repository includes HFGT toolbox input XML files; the version of the HFGT toolbox used to generate system tensors; output HDF5 files for each simulation; Julia scripts used to instantiate, simulate, and visualize the models; and the scripts used to generate all figures presented in this paper. p. 3 Definition 2 – Ontology [21, 13]: An explicit specification of
a conceptualization that defines the concepts, relationships, and
constraints used to represent knowledge within a particular domain. Paper Objective
This tutorial introduces Hetero-functional Graph Theory
(HFGT) as a methodology for constructing ontology-based
models of environmental systems of systems. This specialization, however, presents a challenge when environmental problems span multiple interacting systems. Much
of the structural knowledge embedded within existing process
models remains implicit within parameterizations, calibration
choices, model couplings, or software implementations rather
than being represented explicitly in a form that can be readily
shared or reused across disciplines [10, 9, 14, 15, 16].Where the paper says it
Where the paper says it
Green — the paper says it, and the page is printed. Red — no passage carries it, so treat that sentence as unverified. The same check runs on your own text: check a summary against its source, $5.
Improvements for AI systems
-
Model form before mathematics: AI systems can first model system form and function using SysML-based reference architectures to ensure structural integrity before deriving mathematical formulations.
-
Separation of structure and physics: The improved AI system can maintain a separation between the Engineering System Net, which captures topology and conservation laws, and application-specific device models, allowing for independent refinement of physical processes.
-
Scalable modeling through incremental development: AI systems can develop complex environmental models incrementally by adding new model instances to a common ontology rather than requiring complete changes to the underlying modeling methodology.
-
Unified mathematical framework: The system can use the Hetero-functional Graph Theory and Hetero-functional Network Minimum Cost Flow formulation to translate conceptual descriptions into executable mathematical models, ensuring consistency across disciplines.
-
Data-driven simulation validation: AI systems can automatically generate hetero-functional incidence tensors from system descriptions, which are then used in Julia-based simulators to validate results against continuous time dynamics and discrete time steps.
Abstract
This introductory overview describes a unified modeling framework for representing interconnected environmental systems using Model-Based Systems Engineering and Hetero-functional Graph Theory. Existing environmental models encode rich structural and process-level information, yet much of this knowledge remains implicit within domain-specific formulations. This paper demonstrates how such information can be formalized into a consistent ontology that explicitly represents system structure, function, and interconnections. Using illustrative hydrological examples, the tutorial illustrates how conventional process models can be translated into a scalable system-of-systems representation. Agricultural nitrogen applied to land is represented explicitly as the point at which human decisions enter the hydrological system, illustrating how models of human systems can be coupled to hydrological models within the same representation. These examples highlight how embedded relationships in existing models can be made explicit and computationally tractable. Using simplified examples, the overview provides a foundation for integrating heterogeneous models across environmental and engineered domains. More broadly, this work offers a step-by-step introduction to a generalizable modeling framework for complex systems.
Sources
- The Hetero-functional Graph Theory Toolbox
- A Hetero-functional Graph State Estimator for Watershed Systems: Application to the Chesapeake Bay
Related papers
- Industrial Load Modeling and Optimization for Market-Based Interaction with Power Systems
- Exact and Approximate Convex Reformulation of Linear Stochastic Optimal Control with Chance Constraints
- Certified Set Convergence for Piecewise Affine Systems
- Symbolic Control for Stochastic Systems via Finite Parity Games
- Multiobjective optimization-based design and dispatch of islanded, hybrid microgrids for remote, off-grid communities in sub-Saharan Africa
- Physiology-informed layered sensing for intelligent human-exoskeleton interaction