Inversion of Reflected Travel Time Curve Using a Continuous Genetic Algorithm

Summary Genetic algorithm (GA) uses a directed random search technique applied on global optimization of functions. In this paper we show the application of continuous GA to the inversion of reflected travel time curve. In this algorithm a number of sub-populations are generated to widely explore the search space. Then a new population is created in the neighborhood of best result extracted from these sub-populations. The genetic operators (selection, crossover and mutation) are applied on each population. We have also used an exponentially decreasing mutation probability for better exploration of the search space. This modified GA, when applied on noise free and noise corrupted synthetic data indicates a good convergence to the optimum results.