Many economic models are naturally formulated as stochastic dynamic programming problems, since this approach allows for optimization over time in an uncertain environment. A problem with direct application of stochastic dynamic models is that minimally realistic models that allow for interactions and feedbacks between equations lead to extremely complex computational problems. These problems have until recently impeded the developement and application of stochastic dynamic models. Recent research in computational economics and simultaneous advances in computing technology have made it more feasible to astimate realistically complex stochastic dynamic models. New numeric methods such as simulated annealing and simulation methods of estimation allow for estimation and analysis of previously intractible models. These new methods have very wide applicability an dopen the field for a new level of realism and complexity in economic modeling. The research project proposes to develop methods for estimation and analysis of stochastic dynamic models. A subgroup of researchers will focus on econometric methods, while another subgroup will focus on application of existin tools for stochastic dynamic models to the analysis. The goal of the project is to contribute both methodological advances as well as high quality applications to specific problems.
|Effective start/end date||21/12/97 → 21/12/00|