Exploiting Symmetry in High-Dimensional Dynamic Programming
We propose a new method for solving high-dimensional dynamic programming problems and recursive competitive equilibria with a large (but finite) number of heterogeneous agents using deep learning. The "curse of dimensionality" is avoided due to four complementary techniques: (1) exploiting symmetry in the approximate law of motion and the value function; (2) constructing a concentration of measure to calculate high-dimensional expectations using a single Monte Carlo draw from the distribution of idiosyncratic shocks; (3) sampling methods to ensure the model fits along manifolds of interest; and (4) selecting the most generalizable over-parameterized deep learning approximation without calculating the stationary distribution or applying a transversality condition. As an application, we solve a global solution of a multi-firm version of the classic Lucas and Prescott (1971) model of "investment under uncertainty." First, we compare the solution against a linear-quadratic Gaussian version for validation and benchmarking. Next, we solve nonlinear versions with aggregate shocks. Finally, we describe how our approach applies to a large class of models in economics.
We would like to thank Benjamin Bloem-Reddy, Samira Ebrahimi Kahou, Kevin Leyton-Brown, Yaniv Plan, Kevin Song, and Vincent Michalski for useful comments. The views expressed herein are those of the authors and do not necessarily reflect the views of the National Bureau of Economic Research.