VERTICAL SEISMIC PROFILE INVERSION WITH GENETIC ALGORITHMS

We describe a new genetic-algorithm (GA) inversion technique and apply it to a vertical seismic profile (VSP) inversion problem where the go al is to recover slowness and impedance profiles. Our algorithm consis ts of one long (300 generations) and a series of short (50 generations ) GA runs. After each run the problem is reparametrized using the matr ix of partial second derivatives in a neighbourhood of the best model found in that run. The new unknown parameters are more nearly independ ent than the old ones, thus successive GA runs find better models fast er, in accordance with the principle of meaningful building blocks. Fo r the VSP problem we use a fitness function that is the weighted sum o f the cross-correlation between the measured and synthetic total wavef ields and the cross-correlation between the measured and synthetic upg oing wavefields. The measured upgoing wavefield is obtained from the m easured total wavefield by f-k filtering.