Global optimization algorithms are generally computationally intensive processes, where a significant amount of time is required to generate a solution. Therefore, methods to improve the efficiency are desired for problems that require their use. In this study we present a hybrid method to improve the convergence rate of an AVO inversion through implementation of a trace integration method followed by a simulated annealing optimization. The method eliminates the need for an additional model parameter to represent the layering solution and reduces the compute time of the global optimization by generation of an initial model that is close to the final solution.
View full article as PDF (0.17 Mb)