Optimal designs for probability-based optimality, parameter estimation and model discrimination

http://dx.doi.org/10.1016/j.joems.2017.01.002

Abstract

In this paper, a new compound optimality criterion will be introduced. This criterion called PDKLoptimality. The proposed criterion aimed to introduce designs satisfy maximum probability of success,
efficient parameter estimation and true model. An equivalence theorem is stated and proved for PDKLoptimality criterion.