HICSS minitrack session on self-organized criticality

Abstract

Complex Systems with characteristics similar to SelfOrganized Criticality (SOC) or Highly-Optimized Tolerance (HOT) exhibit failure events of all sizes, typically with a frequency of occurrence described by a heavy-tailed distribution. Some large scale engineered or physical systems, such as the North American electric power grid, have failure statistics consistent with SOC and HOT systems. This suggests that the risks of large-scale failure of these systems should be analyzed using ideas from complex systems. The challenge is to work out which factors in the dyamics, topology and interactions in these systems contribute strongly to the complex behavior and then to construct models capturing these effects so that they can studied with a view to understanding and mitigating catastrophic failures. We pose some of the interesting issues as questions: • What are the salient system characteristics causing heavy tailed distributions of failure events? For example, what emphases should be placed on failure modes, engineering optimization, system loading, network topology, nonlinear dynamics, types of interactions between elements, and social and economic forces? What time scales are involved? • How does engineering of the system in response to societal requirements affect these systems? Is engineering itself part of the dynamics shaping these systems?