We describe a method of nonlinear inversion of travel-time tomography data based on a global optimisation method called Simulated Annealing (SA). The simulated annealing has recently been found to be computationally efficient in several geophysical parameter estimation problems. The solution is independent of the initial model because local minima are avoided during the process by convenient up-hill down-hill exploration of the model space. Preliminary results using synthetic model show how a moderately complex structure can be obtained by SA.
Calculus based algebraic algorithms like Singular Value Decomposition, Ridge Regression, Weighted Ridge Regression, etc. were used in the inversion of resistivity data till late 80's. With the advent of high speed computers having higher disk space and parallel processing environment, several global optimization techniques (simulated evolution, simulated annealing, very fast simulated annealing, etc.) are adapted to solve various inherent problems in the inversion of geophysical data. In the present paper an attempt is made to use Evolutionary Programming (EP) technique for the estimation of layer parameters from one-dimensional Vertical Electrical Sounding (VES) data. The algorithm is coded in FORTRAN 77 and tested on both synthetic and actual field data. The results, compared with those obtained by Singular Value Decomposition and borehole lithology, prove the efficiency and robustness of the EP approach.
JavaScript jest wyłączony w Twojej przeglądarce internetowej. Włącz go, a następnie odśwież stronę, aby móc w pełni z niej korzystać.