Moment functions for the solution of the Cauchy problem for a first-order linear differential equation with random coefficients (Q5930599)

The authors consider the Cauchy problem for the first-order linear differential equation with random coefficients \[ \partial u(t,x,y)/\partial t= \varepsilon_1(t)\partial u(t,x,y)/\partial x+ \varepsilon_2(t)\partial u(t,x,y)/\partial y+ f(t,x,y),\tag{1} \] \[ u(t_0, x,y)= u_0(x,y),\tag{2} \] where \(t\in [t_0,t]= T\subset\mathbb{R}\), \((x,y)\in \mathbb{R}^2\), \(\varepsilon_1\) and \(\varepsilon_2\) are random processes, \(f: T\times \mathbb{R}^2\to\mathbb{R}\) is a random processes, and \(u_0: \mathbb{R}^2\to \mathbb{R}\) is a random process independent of \(\varepsilon_1\), \(\varepsilon_2\), and \(f\). For application, it is important to know the moment function of the solution of problem (1), (2). The authors find the first two moment functions of the solutions. The authors assume that the random processes \(\varepsilon_1\), \(\varepsilon_2\), and \(f\) are given via the characteristic functional \[ \varphi(v_1,v_2, w)= Me(v_1,v_2, w), \] \[ e(v_1,v_2, w)= \exp\Biggl(i\int_T [\varepsilon_1(t) v_1(t)+ \varepsilon_2(t) v_2(t)] dt+ i\int_T \int_{\mathbb{R}} \int_{\mathbb{R}} f(t,x,y) w(t,x,y) dy dx dt\Biggr), \] where \(M\) is the expectation with respect to the distribution function of the processes \(\varepsilon_1\), \(\varepsilon_2\), and \(f\). To solve the problem, the authors need the solution of an equation with the variational derivatives.