Relation between Lacunarity, Correlation dimension and Euclidean dimension of Systems

Abstract: Lacunarity is a measure often used to quantify the lack of translational invariance present in fractals and multifractal systems. The generalised dimensions, specially the first three, are also often used to describe various aspects of mass distribution in such systems. In this work we establish that the graph (\textit{lacunarity curve}) depicting the variation of lacunarity with scaling size, is non-linear in multifractal systems. We propose a generalised relation between the Euclidean dimension, the Correlation Dimension and the lacunarity of a system that lacks translational invariance, through the slope of the lacunarity curve. Starting from the basic definitions of these measures and using statistical mechanics, we track the standard algorithms- the box counting algorithm for the determination of the generalised dimensions, and the gliding box algorithm for lacunarity, to establish this relation. We go on to validate the relation on six systems, three of which are deterministically determined, while three others are real. Our examples span 2- and 3- dimensional systems, and euclidean, monofractal and multifractal geometries. We first determine the lacunarity of these systems using the gliding box algorithm. In each of the six cases studied, the euclidean dimension, the correlation dimension in case of multifractals, and the lacunarity of the system, together, yield a value of the slope $S$ of the lacunarity curve at any length scale. The predicted $S$ value matches the slope as determined from the actual plot of the lacunarity curves at the corresponding length scales. This establishes that the relation holds for systems of any geometry or dimension.