Application of Genetic Algorithm in Inverse Problem of Welltesting Interpretation of Triple Media Reservoirs

摘要

It is obviously impossible to match test data of triple media reservoirs by hand because of too many parameters to be interpretated. The different parameters of well-testing interpretation have bad relativity and the well-testing interpretation is a non-linear inversion problem with multi-parameters. Genetic Algorithms which are developed recently have advantage of global convergence. It is auto-adapted and non-linear optimum algorithm without gradient information which is very fit for the well-testing interpretation of triple media reservoirs. It is very successful for the auto-match of well-testing interpretation of triple media reservoirs by use of Genetic Algorithms which improves the matching speed and precision between theoretical pressure and the observed pressure.