Content uploaded by Eric Tatara
Author content
All content in this area was uploaded by Eric Tatara on May 06, 2015
Content may be subject to copyright.
The Simphony Integrated Simulation Framework for Infrastructure Interdependency Modeling
Eric Tatara and Michael J. North
This paper presents an introductory tutorial and illustration of the modeling capabilities of the
Repast Simphony simulation framework using an agent-based model of regional natural gas and
electric power system interdependencies. The natural gas transmission and distribution and
electrical power companies are modeled as social agent organizations. The goal of the modeling
exercise is to simulate impacts on the delivery of natural gas to the affected geographical regions
and the effects of various policy models in the agent-based company models. Additionally, the
estimate of cascading effects of natural gas loss on associated electric power infrastructure is
investigated.
The physical layers in the simulation consists of a natural gas transmission network and DC
electric transmission network.
The natural gas transmission model consists of a network of interconnected links and nodes,
where the nodes function as delivery, receipt and/or pipeline termination points and the links
function as gas pipelines that transport natural gas between nodes. Supplies, demands and
compressor fuel usage values and compressor and pressure regulator characteristics may be
modified to observe the effects on the network.
The DC electric network model considers a balance of demand and generation given the
transmission topology. The nodes in the electric network represent generators and load points,
while the links function as electrical transmission lines. The electric network model permits
sensitivity analysis of line outages and shifts in generation. The two networks are connected via
links between the natural gas network and gas-fired electric power plants (generators) in the
electrical network. Each network is fairly large scale and contain several thousand nodes and
links each.
The social agent layers consists of both regional and local gas transmission companies as well as
local electrical power companies.
The regional natural gas transmission companies control the portion of the network between the
gas supply and production sources. The local gas distribution company is responsible for
delivery of the gas between the local city gates and the customer base (commercial and
residential) Both natural gas companies share a common internal structure consisting of several
groups such as operations, engineering and supply control. Each company group is modeled as a
collection of different agent types, each with a unique type of behavior suited to their respective
tasks. Certain agents act as a liaison between the companies and between the transmission and
local distribution companies.
The social and physical agent models were developed in Repast [ROAD 2005], a widely used,
free, and open source, agent-based modeling and simulation toolkit with three released
platforms, namely Repast for Java, Repast for the Microsoft .NET framework, and Repast for
Python Scripting (North et al 2005). Repast Simphony (Repast S) extends the Repast portfolio by
offering a new approach to simulation development and execution (North et al 2006). Simphony
goes beyond current model integration architectures that simultaneously require users with broad
and deep knowledge in the application area, modeling, and software engineering. This broad
skill set rarely exists within a single organization and is even less likely to be integrated into a
working team. Simphony brings model integration tools directly to domain expert analysts (e.g.,
infrastructure specialists) without the need of software engineers, allowing users to develop
integrated simulation modeling applications much more efficiently and economically than has
been previously possible.
Simphony provides drag-and-drop user interfaces for: selecting from a repository of model
components and tools (e.g., geographical information systems (GIS), visualization tools,
statistical tools, and databases); creating new models using flow diagrams and agent-based
model templates; assigning data sources such as GIS data, relational databases or real-time data
feeds such as sensors; defining how components should interact; and controlling how the
simulation is executed. The Simphony team is also investigating the use of grid computing
technologies to provide large-scale simulation capabilities and distributed collaborative
development of integrated national security applications.
To insure that Simphony directly addresses analyst’s needs, its design is based on a “Use Case”
approach where example applications are presented to end users and progressively refined using
their suggestions. Several use cases are being developed including infrastructure
interdependency; natural gas disruption fast turn around analysis; energy process and market
analysis; supply network infrastructures; and manufacturing and industrial processes.
[ROAD 2005] Repast, ROAD (Repast Organization for Architecture and Design), 2005, Repast
Home Page, Chicago, IL; available at http://repast.sourceforge.net.
[North, Howe, Collier, & Vos, 2005b] North, M.J., T.R. Howe, N.T. Collier, and J.R. Vos, 2005,
“Repast Simphony Runtime System,” in C.M. Macal, M.J. North, and D. Sallach (eds.),
Proceedings of the Agent 2005 Conference on Generative Social Processes, Models, and
Mechanisms, ANL/DIS-06-1, co- sponsored by Argonne National Laboratory and The
University of Chicago, Oct. 13–15.
[North, Collier, & Vos, 2006] North, M.J., N.T. Collier, and J.R. Vos, 2006, “Experiences
Creating Three Implementations of the Repast Agent Modeling Toolkit,” ACM Transactions on
Modeling and Computer Simulation 16(1):125, ACM (January): New York, NY.
