Weighting Schemes and Objective Functions in Sub-Surface Flow and Transport Parameter Identification via Inverse Simulation
Carey, Martha M.
We propose a number of different weighting schemes (piecewise linear, quadratic, sine and narrow-base sine) based on data magnitude for least-squares minimization of concentration data for parameter estimation for groundwater flow and contaminant transport. We plot the resulting objective functions along lines of varying single distributed parameters for a small simulated data set. We observe that weighting can remove problems of spurious local minima and non-convexity, and that narrow-base sine weights produce good results for varying sample strategies and underlying structures.
Citable link to this pagehttps://hdl.handle.net/1911/101870
MetadataShow full item record
- CAAM Technical Reports