I.M. Sokolov,A.V. Chechkin,J. Klafter
Abstract:Truncated Levy flights are stochastic processes which display a crossover from a heavy-tailed Levy behavior to a faster decaying probability distribution function (pdf). Putting less weight on long flights overcomes the divergence of the Levy distribution second moment. We introduce a fractional generalization of the diffusion equation, whose solution defines a process in which a Levy flight of exponent alpha is truncated by a power-law of exponent 5 - alpha. A closed form for the characteristic function of the process is derived. The pdf of the displacement slowly converges to a Gaussian in its central part showing however a power law far tail. Possible applications are discussed.
What problem does this paper attempt to address?
The problem that this paper attempts to solve is about the mathematical description of Truncated Lévy flights. Specifically, the authors introduced a new form of fractional - order diffusion equation to describe Lévy processes with power - law truncation characteristics. This process exhibits the characteristics of Lévy flights on a short - time scale, that is, it has a heavy - tailed probability distribution function (PDF), but on a long - time scale, it will gradually converge to a Gaussian distribution, although this convergence may be very slow.
### Main contributions of the paper
1. **Proposing a new fractional - order diffusion equation**:
- The authors proposed a fractional - order diffusion equation in the form of:
\[
\left(1 - C_\alpha \frac{\partial^{2-\alpha}}{\partial |x|^{2-\alpha}}\right) \frac{\partial p(x,t)}{\partial t} = D \frac{\partial^2 p(x,t)}{\partial x^2}
\]
where \( D \) is the diffusion coefficient that controls the long - time asymptotic behavior, \( C_\alpha=\frac{D}{K_\alpha} \) is the proportional factor that controls the intermediate - time Lévy - like behavior, and the dimension of \( C_\alpha \) is \([C_\alpha]=[L^{2-\alpha}]\).
2. **Analytical solutions and eigenfunctions**:
- Through Fourier transform, the authors derived the eigenfunction of this equation:
\[
\left(1 + C_\alpha |k|^{2-\alpha}\right) \frac{\partial f(k,t)}{\partial t} = -D k^2 f(k,t)
\]
- The specific form of the eigenfunction is:
\[
f(k,t)=\exp\left(-\frac{D k^2}{1 + C_\alpha |k|^{2-\alpha}}t\right)
\]
3. **Properties of the probability density function**:
- The authors proved that the solution of this equation is a valid probability density function (PDF) and discussed its main properties. In particular, for large \( x \), the asymptotic behavior of the PDF is:
\[
p(x,t)\simeq\frac{\Gamma(5-\alpha)\sin(\pi\alpha/2)}{\pi D C_\alpha t}\frac{1}{x^{5-\alpha}}, \quad x\rightarrow\infty
\]
- This shows that the Lévy distribution is truncated by a steeper power - law function, and the truncation power is \( 5-\alpha \).
4. **Application prospects**:
- The authors discussed the potential applications of this model in economics, especially in the modeling of commodity price behavior. For example, the cumulative distribution function of cotton prices may correspond to the power - law behavior \( 1 - F(x)\propto x^{-\alpha} \) in the intermediate part and the power - law behavior \( 1 - F(x)\propto x^{-\beta} \) in the far - tail part, where \( \alpha = 1.7 \) and \( \beta\approx3 \). This model can well describe this situation, giving \( \beta = 3.3 \).
### Summary
This paper successfully described the Lévy flight process with power - law truncation characteristics by introducing a new fractional - order diffusion equation. This model is not only of great theoretical significance, but also shows great potential in practical applications, especially in the field of economics.