Roots

The roots of this initiative grew in the fertile soil of the National Infrastructure Simulation and Analysis Center (NISAC) program at Sandia National Laboratories (Sandia). The USA Patriot Act of October 2001 tasked NISAC "to serve as a source of national competence to address critical infrastructure protection and continuity through support for activities related to counter terrorism, threat assessment and risk mitigation." This task specified modeling, simulation and analysis of critical infrastructure systems to enhance understanding of their large-scale complexity and facilitate system modifications to mitigate threats and enhance the stability of critical infrastructures. In short, the mandate is to integrate modeling, simulation, and analysis into national infrastructure and asset protection planning and decision support activities.

In 2002 we formed the Advanced Methods and Techniques Investigation group within NISAC to filter concepts from the Complex Systems literature that cross disciplines and to identify study/modeling approaches that could be of direct use to NISAC. We also worked to identify common unresolved issues pertaining to the application of Complex Adaptive Systems modeling to real-world situations and defined our R&D effort for Complex, Interdependent Adaptive Infrastructure to resolve the issues of greatest importance to NISAC.

Our course of study focused on an examination of the scientific literature in Complexity Science. This literature is extremely diverse and is rapidly encompassing almost every field of scientific investigation. The sorting of this literature to first identify the critical set of topics, and then to find the critical papers in each, took a significant effort (and made all our heads spin from time to time!). We covered these critical papers in a series of Sessions where participants would first read the assigned papers and then come together to discuss them in two-hour periods. Topics covered included:

Background / Overview

Self-Organized Criticality…From Physical to Social Sciences

Beyond SOC… and what about model validation?

Introduction to Complex Networks

Adaptive Networks

In-depth Graph Theory as relevant to Complex Networks

Starting to dig into Social Behavior and Agent-Based Models

Modeling Social Behavior

Modeling Financial Behavior

Modeling Policy

Artificial Chemistry

Our course of study is included as Appendix A in the project report Defining Research and Development Directions for Modeling and Simulation of Complex, Interdependent Adaptive Infrastructure.

During our research, a description of infrastructures naturally emerged that is similar to the way in which complex systems have been described in physics, biology, economics and other diverse areas where Complexity Science has provided fruitful insights. We developed a representation of the elements and interactions in a complex adaptive network (illustrated below) which became our conceptual lens for system investigation.

In most real systems, the rules, parameters and connections that make up a network change over time. A network's current structure and dynamics together determine its performance. These elements change over time, and those changes may depend in some way on the system's overall performance. In this feedback process, either local or global performance measures guide the modification of the local properties of the system. To understand this larger process, we must understand which elements of the network (e.g. structure, transition rules, propagation rules) are subject to change and how the performance of the system influences these changes.

Over the next two years, we applied our understanding to congestive failure in electric power and large banking systems, the next steps along our evolution of the CASoS Initiative at Sandia.