On the global existence of Bohmian mechanics (Q1907969)

This paper proves that the equation of D. Bohm's mechanics have solutions free of singularities for almost all solution curves in an appropriate sense. Mathematically, the problem is to find a wave function \(\psi(Q)\) and particle locations \(Q_j(t)\) such that \(\psi\) solves an \(N\)-particle Schrödinger equation with potential \(V(Q)\), and the \(Q_j\) evolve according to \(dQ_j/dt= \hbar m_j^{- 1}\text{Im}(\nabla_j\psi(Q)/\psi(Q))\). The Schrödinger equation is independent of the particle motion. Admissible interaction potentials include those of the form \(r^{- \alpha}\), \(\alpha< 2\), or more generally those in \(L^{3/2}+ L^\infty\) (a more general statement is given in the paper). The result requires essentially that \(\psi\) be \(C^\infty(H)\), and that \(|\nabla\psi_t|^2\) be integrable over all finite time intervals. Comparison with the classical case, where several types of singularities do occur, is detailed. Brief remarks on the applicability of the hydrodynamic interpretation of the probability density (Madelung) are included. The statement about ``almost all'' data is defined with the help of a stochastic process with death associated with the particle trajectories, the measure being defined by the initial value of \(\psi\).