Stats4Grads: Solving Highly Difficult Policy Decision Problems for Complex Systems with Computer Simulators and Sequential Emulation
5 November 2008 14:15 in CM105
We discuss combining expert knowledge and computer simulators in order to provide decision support for policy makers in complex decision problems. We allow for future states of the complex system of interest to be viewed after initial policy is made, and for those states to influence revision of policy. The potential for future observations and policy intervention impacts heavily on optimal policy for today and this is carefully handled. We write down the decision tree for this policy problem, and show how model-based forecasts of the system under different policy scenarios can be used in a sequential emulation of the tree. We show how to emulate a tight upper bound on the expected loss of a given policy so that it can be shown to a decision maker who may then use it to make policy, or to refine a search for optimal policy by observing which policies perform so poorly that they are not worth considering.
Contact firstname.lastname@example.org for more information
See the Stats4Grads page for more details about this series.