Two-dimensional probability distribution of the solution to the random Burgers-Riemann problem (Q6111908)

The paper studies the solution to the following equation \[ \partial_t u(x, t) + u(x, t)\partial_x u(x, t) = 0 \] with \(u(x, 0) = u_l\) or \(u_r\) according as \(x<0 \) or \(x>0\). The probability distributions of \((u_l, u_r)\) are assumed to be absolutely continuous and known. The marginal probability distribution of \(u(x, t)\) is known in terms of the density of \((u_l, u_r)\). While the previous derivations of this marginal probability distribution used somewhat standard probabilistic techniques, the author chooses to provide an alternative proof using the formalism of Dirac's delta function and its associated calculus. A key component of the proof seems to be the idea that one can view the probability density function itself as an average (of impulses), as suggested by the identity \[ f_X(a) = E(\delta(a - X)) \] for a random variable \(X\) with probability density \(f_X(\cdot)\), and \(E\) denoting the expectation operator. The primary contribution of the paper is the joint distribution of \((u(x_1, t_1), u(x_2, t_2))\), which the author derives using the Dirac's delta function formalism. While I believe the same calculations could be done using standard probabilistic tools, the convenience of the Dirac's delta function will certainly appeal to the physics and applied mathematics community.