Dean S. Oliver
Principal Researcher
Uni Centre for Integrated Petroleum Research
Uni Research
Bergen, Norway

Dean Oliver

E-mail address:

Office address:
Uni CIPR (Centre for Integrated Petroleum Research)
Realfagbygget, 4th floor
Allégaten 41
N-5007 Bergen


Dean Oliver is a principal researcher at the Uni Centre for Integrated Petroleum Research. Before moving to Norway, he was the Mewbourne Chair Professor in the Mewbourne School of Petroleum and Geological Engineering at the University of Oklahoma and the Director of the OU Center for Ensemble Methods. He holds a BS degree in Physics and a PhD in Geophysics. He was Director of the Mewbourne School from 2002 to 2006. Prior to joining The University of Oklahoma, he was a professor in the Petroleum Engineering department at The University of Tulsa for six years, the last year as Chairman. He worked seventeen years for Chevron as a research geophysicist, a staff reservoir engineer for Chevron USA and for Saudi Aramco, and as a research scientist in reservoir characterization. He has received best paper of the year awards in two journals and received the SPE Reservoir Description and Dynamics award in 2004. He was awarded Distinguished Member status in SPE in 2008, the SPE Distinguished Service Award in 2010, and Honorary Membership in 2017. He was the Executive Editor of SPE Journal 2005--2009 and the Editor-in-Chief of SPE journals 2013-2017. He serves on the Board of Directors of Resoptima AS. His research interests are in inverse theory, reservoir characterization, uncertainty quantification, and optimization.

More details can be found in Dean Oliver's cv.

Research projects

History matching of geologic facies. Current research projects focus on the use of the ensemble Kalman filter and smoothers for data assimilation and control of petroleum reservoirs. Several problems are of particular interest.

  1. Application of ensemble methods to highly nonlinear data-model relationships such as might be found in multiphase flow in porous media. Because of the nonlinearity of the relationships, the pdf for parameters can be multimodal in parts of the reservoir. We are pursuing several approaches, one hybrid approach that requires a combination of ensemble methods with more traditional gradients, and another that weighting of samples.
  2. Application of ensemble methods to highly non-Gaussian model variables such as might occur in reservoirs with geologic facies. Previous work on history matching of 2D facies boundaries is being extended to 3D reservoirs.
  3. Application of ensemble methods to reservoir problems with large numbers of correlated state variables. We are particularly interested in compositional modeling.
  4. Robust optimization of reservoir production using ensemble methods both for data assimilation and for optimal control.

Dean S. Oliver <>