Iteratively constructive sequential design of experiments and surveyswith nonlinear parameter-data relationships

摘要

In experimental design, the main aim is to minimize postexperimental uncertaintyon parameters by maximizing relevant information collected in a data set. Using anentropy-based method constructed on a Bayesian framework, it is possible todesign experiments for highly nonlinear problems. However, the method iscomputationally infeasible for design spaces with even a few dimensions. We introduce aniteratively constructive method that reduces the computational demand by introducingone new datum at a time for the design. The method reduces the multidimensional designspace to a single-dimensional space at each iteration by fixing the experimental setupof the previous iteration. Both a synthetic experiment using a highly nonlinearparameter-data relationship and a seismic amplitude versus offset (AVO) experiment areused to illustrate that the results produced by the iteratively constructive methodclosely match the results of a global design method at a fraction of the computational cost.This work thus extends the class of iterative design methods to nonlinear problemsand makes fully nonlinear design methods applicable to higher dimensionalreal-world problems.