INVERSION OF POTENTIAL-FIELD DATA BY GENETIC ALGORITHMS

We present a genetic algorithm that simultaneously generates a large n umber of different solutions to various potential field inverse proble ms. It is shown that in simple cases a satisfactory description of the ambiguity domain inherent in potential field problems can be efficien tly obtained by a simple analysis of the ensemble of solutions. From t his analysis we can also obtain information about the expected bounds on the unknown parameters as well as a measure of the reliability of t he final solution that cannot be recovered with local optimization met hods. We discuss how the algorithm can be modified to address large di mensional problems. This can be achieved by the use of a 'pseudo-subsp ace method', whereby problems of high dimensionality can be globally o ptimized by progressively increasing the complexity and dimensionality of the problem as well as by subdividing the overall calculation doma in into a number of small subdomains. The effectiveness and flexibilit y of the method is shown on a range of different potential field inver se problems, both in 2D and 3D, on synthetic and field data.