Exact Computational Analyses for Adaptive Designs

We show how to compute optimal designs and exact analyses of allocation rules for various sequential allocation problems. The problems we have solved include parameter estimation in an industrial scenario, and testing in a clinical trial. Our computational approach incorporates backward induction, dynamic programming, and a new technique of forward induction. By utilizing efficient algorithms and careful implementation, we are able to determine exact solutions to practical problems previously approached only through simulation or approximation. Keywords: constrained dynamic programming, forward induction, backward induction, bayesian design, multiple criteria, clinical trials 1 Introduction We discuss the exact analyses of sequential allocation rules for a variety of applications. Our interests include both estimation and testing problems, and we have examined allocation rules for selected applications in medicine and industry. In the former, we study the design and evaluation of opt..