For the equation of wave propagation in the half-space R 2 + = {(x, y) ∈ R 2 | y > 0} we consider the problem of determining the speed of wave propagation that depends only on the variable y and the shape of a point impulse source on the boundary of the half-space. We show that, under some assumptions on the shape of the source and the structure of the medium, both unknown functions of one variable are uniquely determined by the displacements of boundary points of the medium. We estimate stability of a solution to the problem.
We consider the differential equation in the domainwith c = c(y) some positive function of class C 2 (R + ). Suppose that u(x, y, t) satisfies in addition to (1.1) the following initial and boundary conditions:Given c(y) and f (t), the problem (1.1), (1.2) is well-posed and determines a function u(x, y, t), with compact support for each finite t. In applications, for instance in geophysics, there is some interest in the problem of determining the structure of some medium, in this case the function c(y), from the measured displacements of points of the medium on the boundaryThis problem is called the inverse dynamical acoustic problem; it was considered for a known function f (t) in several articles; for instance, see [1][2][3][4][5][6]. Usually the Dirac delta function δ(t) or some regular function with finite discontinuity at t = 0 played the role of f (t). These are the assumptions on the shape of a source under which a few uniqueness and stability theorems were established for the inverse problem and some numerical algorithms were justified for solving the problem.In this article we consider a more general problem of determining the two functions c(y) and f (t) from the data (1.3). Such formulation is motivated by the circumstance that, physically, the boundary condition (1.2) models a point explosion at the boundary of the domain, in which case the function f (t) cannot be measured directly. To determine it we could only use observations of the oscillating boundary of the domain. Gerver considered [7] a similarly formulated problem for the one-dimensional wave equation corresponding to the case that the solution to the problem is independent of x; for instance, if we eliminate the delta function from the boundary condition (1.2). However, in this case the information