Feasibility of applying the dynamic programming technique to optimally controlling medium to large scale macroeconomic models. A survey article (Q1097826)

We have limited the scope of our study to consider medium to large scale linear and nonlinear macroeconometric models. The fundamentals of dynamic programming have been discussed in three sections. The feasibility of computational algorithms to derive the optimal control solution from medium to large scale economic models has also been introduced. Among the various techniques listed in the paper, which have been proposed to reduce the dimensionality difficulties in dynamic programming and are discussed elsewhere, we will discuss a procedure which decomposes the state space and policy space into different regions and then performs dynamic programming calculations over the whole state and control space by calculating the optimal control solution over each region separately. As far as the medium to large scale nonlinear economic models are concerned, this decomposition technique is the most promising one aimed at reducing the computer memory requirement in applying dynamic programming algorithms.