Bayesian Decision-Theoretic Design of Experiments Under an Alternative Model

Traditionally Bayesian decision-theoretic design of experiments proceeds by choosing a to minimise expectation given loss function over the space all designs. The encapsulates aim experiment, and is taken with respect joint distribution unknown quantities implied statistical model that will be fitted observed responses. In this paper, an extended framework proposed whereby alternative model. Motivation for includes promoting robustness, ensuring computational feasibility allowing realistic prior specification when deriving design. To aid in exploring new framework, asymptotic approximation expected under derived, properties different functions are established. then demonstrated on linear regression versus full-treatment scenario, estimating parameters non-linear discrepancy cubic spline number basis functions.