(The NDSSL website has been reorganized - please visit
The Network Dynamics and Simulation Science Laboratory is pursuing an advanced research and development program for interaction-based modeling, simulation, and associated analysis, experimental design, and decision support tools for understanding large biological, information, social, and technological systems. Extremely detailed, multi-scale computer simulations allow formal and experimental investigation of these systems. The need for such simulations is derived from questions posed by scientists, policy makers, and planners involved with very large complex systems. The simulation applications are underwritten by a theoretical program in discrete mathematics and theoretical computer science that is sustained by more than a decade of experience with the interplay of research and application.
Biological, information, social, and technological systems consist of large numbers of interacting components that together produce a "global system" with properties that are the result of interactions among the representations of the local system elements. Examples of such global systems include urban and regional transportation systems, the United States and world-wide electrical power markets and grids, the Internet, peer-to-peer networks, ad hoc communication and computing systems, bio-signaling systems, ecologies, gene regulatory networks, the public health system, and contagious disease economics. The complicated interactions and interdependencies among the constituent biological, information, social, and technological systems are inherent because the individual components are networked, only interacting with a specified set of components within local time intervals. The interactions can be physical or a matter of convention, such as those imposed by law or social norms, and typically consist of one or more social, biological, or information networks interacting with underlying technological or physical networks. For many reasons, ranging from practical difficulty to the possibility of great harm, simulations are a uniquely capable medium in which representation and analysis can be performed.
A key feature of the Lab's work is the scale and scope of the systems represented. Constructing large simulations of social and technological systems is challenging and novel, since, unlike physical systems, socio-technical systems are affected not only by physical laws but also by human behavior, regulatory agencies, courts, government agencies, and private enterprises. Our interdependent systems simulation suite provides a controlled environment to represent interactions among socio-technical networks, such as extremely large, interdependent urban infrastructure systems consisting of millions of interacting agents. For example, our population and transportation simulation system can represent every individual and their network in an extended urban region, including areas spanning hundreds of square miles and municipalities, at a spatial resolution of meters and a temporal resolution of one second or less. There can be tens of million of individuals each taking roughly five trips every day. Metropolitan Chicago spans approximately two hundred and fifty square miles with more than four hundred municipalities and over ten million inhabitants, and runs on a cluster of approximately hundred nodes to thousands of nodes. The size, scope, and multiple time scales of these representations naturally motivates a high performance computing implementation requiring new engineering design principles.
The mathematical primitive in the interaction-based setting is an
iterated composition of local functions, whereas the traditional
setting is grounded in recursion and grammatical rules of symbol
substitution and rewrite. Moreover, the interaction-based setting
emphasizes what is being computed by interacting systems, rather than
how "hard" it is to compute a given procedure or
class. Interaction-based computational systems are often more like
operating systems than specific algorithms, as they maintain specified
relationships between individual computational elements. They are
infinite-state systems with programs and behavior that evolve in time
as a result of interactions with other sub-systems, and provide the
main source of increasing complexity of very large, ubiquitous
Copyright Virginia Bioinformatics Institute © 2005