Full text loading...
-
Joint Inversion Using Multi-objective Global Optimization Methods
- Publisher: European Association of Geoscientists & Engineers
- Source: Conference Proceedings, 78th EAGE Conference and Exhibition 2016, May 2016, Volume 2016, p.1 - 5
Abstract
The standard deterministic approach to joint inversion is to combine the multiple objectives (data misfits, regularization and joint coupling terms) into a weighted sum (aggregate) and minimize using a descent-based method. This approach has some disadvantages: appropriate weights must be determined for the aggregate, the use of local optimization requires that the objective functions be differentiable and well behaved, and there is the potential for entrapment in local minima. Pareto Multi-Objective Global Optimization (PMOGO) algorithms can overcome these issues. Also, PMOGO algorithms generate a suite of solutions representing the best compromises between the multiple objectives. We have implemented a PMOGO genetic algorithm and applied it to three classes of inverse problems: standard mesh-based problems for which the physical property values in each mesh cell are treated as continuous variables; mesh-based problems in which the cells can only take discrete physical property values corresponding to known or assumed rock units (a lithological inversion); and a fundamentally different type of inversion for which a model comprises wireframe surfaces representing contacts between rock units (essentially a geometry inversion). Joint inversion is greatly simplified for the latter two classes because no additional mathematical coupling measure is required in the objective function.