We analyze the "fractional continuum limit" and its generalization to n dimensions of a self-similar discrete spring model which we introduced recently [21]. Application of Hamilton's (variational) principle determines in rigorous manner a self-similar and as consequence non-local Laplacian operator. In the fractional continuum limit the discrete self-similar Laplacian takes the form of the fractional Laplacian −(−Δ) α 2 with 0 < α < 2. We analyze the fundamental link of fractal vibrational features of the discrete self-similar spring model and the smooth regular ones of the corresponding fractional continuum limit model in n dimensions: We find a characteristic scaling law for the density of normal modes ∼ ω
828of Lévi flights in n-dimensions. In the limit of "large scaled times" ∼ t/r α >> 1 we show that all distributions exhibit the same asymptotically algebraic decay ∼ t −n/α → 0 independent from the initial distribution and spatial position. This universal scaling depends only on the ratio n/α of the dimension n of the physical space and the Lévi parameter α.MSC 2010: 28A80, 35Q84, 35R11, 60E07, 26A33, 82C31 Keywords and phrases: fractional Laplacian, self-similarity, power laws, scale invariance, fractals, Weierstrass-Mandelbrot function, continuum limit, fractional calculus, Fokker-Planck equation, anomalous diffusion, Lévi flights, Lévi (stable) distributions, non-locality
IntroductionSelf-similarity (scaling-invariance) can be found in many problems of physics. There is a need for understanding the dynamics that leads to fractal and self-similarity properties in domains of the physics as varied as flow turbulence and complex materials. In "traditional modelling" the continuum is assumed to have a characteristic length scale which determines the wavelengths where wave fields interact with the microstructure. However, there are numerous materials in nature which are constituted by a scale hierarchy of recurrent microstructure which can be conceived in a good approximation as self-similar. This is true for solids and porous media but also in fluid mechanics, or multicomponent and multiphase flows. As a consequence, there are many areas of modelling that could benefit from a better understanding of the mechanisms underlying the dynamics of objects exhibiting self-similarity properties.In fluid mechanics, synthetic turbulence models are one of many examples of tempering with the input spectrum in order to understand better the physics underlying Lagrangian diffusion [26] and such spectra can be traced to the fractal distributions of velocity accumulation points in the synthetic flow. Fractal approaches are developing rapidly in fluid mechanics. Such approaches consist in either experimentally or numerically interfering with the flow, forcing it through self-similar objects (or through numerical forcing), [36,11,27].On the other hand, there is a large area of research devoted to the so called Fractional Laplacian and a huge number of references exists, employing this operator in various contexts of ...