Measuring Complexity and Critical Behaviour in Spatial Patterns in Afghanistan

13-3-4.jpg
13-3-4.jpg

Measuring Complexity and Critical Behaviour in Spatial Patterns in Afghanistan

9.95

Author(s): Peter Dobias; Kevin Sprague
No pages: 6
Year: 2010
Article ID: 13-3-4
Keywords: operations research, simulation and training
Format: Electronic (PDF)

Add To Cart

Abstract: This paper outlines the use of two quantities, the fractal dimension and symmetropy, the latter a marriage of symmetry and entropy, that have been shown to be effective in describing key aspects of the spatial distribution of systems that display nontrivial spatial patterns and that are regulated by fractal statistics. Motivation for applying these two measures to the geographical distribution of violent incidents in Afghanistan from 2006 to 2009 was provided by previous studies suggesting that this conflict exhibits properties of a near critical system when violent incidents are viewed as a point process. In this paper, the (spatial) fractal dimension and symmetropy are computed for violent incidents in Afghanistan from 2006 to 2009 and noticeable changes are qualitatively linked to known events over the same time period. Overall, it was found that variations in the (spatial) fractal dimension were highly correlated with the data density and that variations in the symmetropy were largely anti-correlated with both. Both measures detected the presence of localized patterns, and symmetropy distinguished a strong pattern formation not related to density, suggesting a possible shift in the conflict dynamics.