Exact Solutions of the Time Derivative Fokker-Planck Equation: A Novel Approach

H. I. Abdel-Gawad,N. H. Sweilam,S. M. AL-Mekhlafi,D. Baleanu
DOI: https://doi.org/10.48550/arXiv.2008.02801
2020-08-07
Abstract:In the present article, an approach to find the exact solution of the fractional Fokker-Planck equation is presented. It is based on transforming it to a system of first-order partial differential equation via Hopf transformation, together with implementing the extended unified method. On the other hand, reduction of the fractional derivatives to non autonomous ordinary derivative. Thus the fractional Fokker-Planck equation is reduced to non autonomous classical ones. Some explicit solutions of the classical, fractional time derivative Fokker-Planck equation, are obtained . It is shown that the solution of the Fokker-Planck equation is bi-Gaussian's. It is found that high friction coefficient plays a significant role in lowering the standard deviation. Further, it is found the fractionality has stronger effect than fractality. It is worthy to mention that the mixture of Gaussian's is a powerful tool in machine learning. Further, when varying the order of the fractional time derivatives, results to slight effects in the probability distribution function. Also, it is shown that the mean and mean square of the velocity vary slowly.
Analysis of PDEs
What problem does this paper attempt to address?
The main problem that this paper attempts to solve is to find the exact solutions of the Fractional Fokker - Planck Equation (FFPE). Specifically, the author proposes a new method to solve this equation by transforming the FFPE into a system of first - order partial differential equations and combining the Extended Unified Method. In addition, the author also studies how to simplify the fractional - order derivative into the derivative form of non - autonomous ordinary differential equations, thereby simplifying the fractional - order Fokker - Planck equation into the classical form. ### Main research contents: 1. **Method introduction**: - A new method is proposed to transform the fractional - order Fokker - Planck equation into a system of first - order partial differential equations through the Hopf transformation. - The Extended Unified Method is used to solve these systems of equations. 2. **Simplification of fractional - order derivatives**: - The fractional - order derivative is simplified into the derivative form of non - autonomous ordinary differential equations, thereby transforming the fractional - order Fokker - Planck equation into the classical form of the Fokker - Planck equation. 3. **Forms of specific solutions**: - Some explicit solutions of the Fokker - Planck equation with fractional - order time derivatives are obtained. - The properties of these solutions are studied, especially their bi - Gaussian distribution characteristics. 4. **Analysis of parameter influences**: - The influence of the friction coefficient on the standard deviation of the distribution function is analyzed, and it is found that a high friction coefficient significantly reduces the standard deviation. - The influence of the order of the fractional - order derivative on the probability distribution function is studied, and it is found that the change in the order of the fractional - order derivative has a slight influence on the tail of the distribution function. 5. **Numerical results**: - The changes in the mean and mean - square values of the distribution function and velocity under different parameters (such as the order of the fractional - order derivative and the friction coefficient) are shown through numerical simulations. ### Main conclusions: - The proposed method can effectively solve the fractional - order Fokker - Planck equation and obtain exact solutions. - The distribution function appears as a mixed - Gaussian distribution. - The friction coefficient plays an important role in reducing the amplitude of the distribution function. - The change in the order of the fractional - order derivative mainly affects the tail of the distribution function, but has little influence on the mean and mean - square values. Through these studies, the paper provides a new theoretical tool and method for understanding and dealing with fractional - order diffusion processes.