Constrained joint inversion of magnetotelluric and seismic data using simulated annealing algorithm

In this paper the continuous model space of the very fast simulated annealing was transformed to a variable discrete. It can be proved by theoretical models that 5 to 10 times calculation workload can be saved. In order to fit the needs of joint inversion, the 'single judge' rule in the error checking of the very fast simulated annealing algorithm was developed to the 'double judge' rule to avoid the judge mistakes made by Metropolis accepting rule. One-dimensional and two-dimensional multi-parameters joint inversion of magnetotelluric and seismic data were put forward under the uneven topography conditions. The example calculation results of the theoretical models and real observed data illustrated that this constraint joint inversion technique was quite effective.