An optimization approach to multi-dimensional time domain acoustic inverse problems

An optimization approach to a multi-dimensional acoustic inverse problem in the time domain is considered. The density and/or the sound speed are reconstructed by minimizing an objective functional. By introducing dual functions and using the Gauss divergence theorem, the gradient of the objective functional is found as an explicit expression. The parameters are then reconstructed by an iterative algorithm (the conjugate gradient method). The reconstruction algorithm is tested with noisy data, and these tests indicate that the algorithm is stable and robust. The computation time for the reconstruction is greatly improved when the analytic gradient is used.