Models based on the theory of Brownian motion are widely used to describe a number of phenomena in various field of modern science. One of the most impressive applications of the theory of diffusion transport are the so-called Brownian motors that simulate the directional motion of particles in a fluctuating periodic potential [1]. These models have found wide application both in description of motor proteins [2] and in the creation of artificial nanoscale devices [3]. On the other hand, in the theory of Brownian motors a number of specific problems appear which can enrich the theory of diffusion transport at all [4]. One of the ways of modeling Brownian motion is based on solving, by computer simulation (molecular dynamics method), the Langevin equation
$$m\ddot{x}(t) = -\zeta \dot{x}(t) - U'(x(t),t) + \xi(t) \tag{1}$$Here -U'(x,t) is the force working on the particle in the potential field U(x,t), $\zeta = \mu^{-1}$ is the viscous friction coefficient (inverse static mobility), $\xi(t)$ is the unbiased $\delta$-correlated Gaussian noise $(\langle \xi(t) \rangle = 0, \langle \xi(t) \xi(t') \rangle = 2k_BT\zeta\delta(t-t'))$, $k_B$ is Boltzmann constant, T is the absolute temperature). This numerical technique makes it possible to simulate conditions close to real experiment. The basic idea of modelling is to integrate numerically the equation of motion with the random (Langevin) force simulated on a computer [5]. Then the required characteristics of the process are calculated by taking the average for a large number of realizations.
In this way, we have revealed the regularities of overdamped $(m \rightarrow 0)$ Brownian motion in the following cases: free diffusion $(U(x,t)=0)$, diffusion under the influence of a stationary homogeneous (tilting) force $(U(x,t)=-Fx)$, diffusion in a periodic potential $(U(x,t)=V(x))$ with $(V(x+L)=V(x))$, and diffusion in a periodic potential under the influence of a stationary and homogeneous force $(U(x,t)=V(x)-Fx)$. For these models, the effective diffusion coefficients and the particle effective mobilities have been calculated as the quantities of foremost interest. It is worthy to note, that under suitable conditions, the last case (diffusion in a tilted periodic potential) can demonstrate the giant acceleration of free diffusion [6, 7] so this effect could be of great interest from Brownian motors point of view as well. The results of statistical tests have been related to the available analytical data (namely, to results of Lifson and Jackson [8], Stratonovich [9] as well as to more resent results [6, 7]). The time dependencies of average particle characteristics as well as histograms of coordinate values and their distribution function have been given. The proposed model can be expanded by the time dependence of the potential energy of the particles (Brownian motor) and inertial effects to be taken into account (so we consider our results as a start position for our future motor experiments).