Brownian motion appears in a number of physical phenomena and is used for development of artificial nanoscale devices [1-3]. One of the most impressive applications of the theory of Brownian motion is the so-called Brownian motors which model the directional motion of particles in fluctuating periodic potentials [1-3]. In the vast majority of cases, this directional motion takes place in a periodic potential profile with a finite life-time [4]; that makes it important to describe Brownian motion just in such states. The main quantity containing information about the diffusion of a particle is the conditional probability, $P(x,t|x_0,0)$. Supposing that a state (a certain potential profile) can decay with the rate $\Gamma$, i.e, the probability of decay in the time interval [t,t+dt] is $\Gamma \exp(-\Gamma t)dt (t \ge 0)$, the probability to find the particle in this state with a finite life-time $\Gamma^{-1}$ is defined as
$$G(x,x_0) = \Gamma \int_0^\infty dt P(x,t|x_0,0) \exp(-\Gamma t).\tag{1}$$For the diffusion of a Brownian particle in a periodic potential V(x) (L is the period), the quantity G(x,y) can be considered as the Laplace representation of the Green's function of the Smoluchowski equation; thus the equation for G(x,y) appears as
$$\left[1 + \Gamma^{-1} \frac{\partial}{\partial x} \hat{J}(x)\right] G(x, y) = \delta(x - y), \qquad \hat{J}(x) = -De^{-\beta V(x)} \frac{\partial}{\partial x} e^{\beta V(x)} = -D \frac{\partial}{\partial x} - \beta D V'(x) \tag{2}$$Here $\beta=(k_BT)^{-1}$ is the inverse thermal energy ($k_B$ is the Boltzmann constant, T is the absolute temperature), D is the diffusion coefficient. Using the conditions of periodicity, of continuity of G-function and the flux discontinuity at point y=x, we have got the system of equations for the coefficients of periodic particular solutions of Eq. (2). For piecewise-linear potentials, with two linear links ([0,l] and [l,L]) and an energy barrier V, experiencing dichotomous fluctuations with the inverse correlation time $\Gamma$, the analytical solution of this system has been found. This result determines the probability density to find the particle, being initially placed at the point y, in the state with a finite life-time $\Gamma^{-1}$ at the point x. The detailed graphical analysis of this solution has been carried out which gave all the information about the diffusion in the potentials considered; it was shown that the behavior of G(x,y) strongly depends on the position y relative to the narrow and wide links of the potential profile as well as on the ratio of the potential amplitude to the thermal energy, $\beta V$. Thus, the function G(x,y) is an effective tool to consider the influence of parameters of the model on the average velocity of a Brownian motor. Using the approach suggested, the average velocity of stochastic Brownian motors, of flashing and rocking as well as of combined types, has been calculated and analyzed. Both analytical results and the results of calculation demonstrate good correspondence with the known results [4,5] in limiting cases.