Improved hierarchical surrogate-assisted evolutionary algorithm with multiscale variable-reduction strategy for large-scale optimization
| Main Author: | Cong Xiao |
|---|---|
| Format: | info software Journal |
| Terbitan: |
, 2021
|
| Subjects: | |
| Online Access: |
https://zenodo.org/record/4663521 |
Daftar Isi:
- Identifying optimal well operating conditions is critical to enhance the hydrocarbon recovery in the petroleum industry. Traditionally, this was addressed by testing numerous design scenarios from integrated physical models and therefore presents computational intractability and and infeasibility in mathematical optimization. Therefore, we propose an innovative derivative-free optimization framework, which considerably reduces computing time from months to days by exploiting state‐of‐the‐art multi-fidelity surrogate modeling technology. Specifically, this paper proposes a novel hierarchical surrogate-assisted evolutionary algorithm based on multiscale variable-reduction strategy for application to the large-scale production optimization step of closed-loop reservoir management where the objective function is the net present value (NPV) of production from a given reservoir. In addition, well-controls are regularized as well using function control method (FCM) and interpolation control method (ICM) in the literature to make evolutionary algorithm much more well-poised and efficient. We also define a transformation functions, e.g., sigmoid function, to make the well control constrains automatically satisfied and therefore ease the implementation of entire optimization framework.