CEM
Cross-Entropy Method (CEM) for Economists (with Daniel Albuquerque, Dina Gat, and Yochai Gat) - The CEM is an optimization method that is derivatives-free and can handle discontinuities, arbitrary constraints, and a large parameter space with ease. Also, it is simple, transparent, and flexible, which makes it a valuable tool for fitting economic models to data. This guide provides insights into its origins and connections to existing methods. We showcase its successful application in estimating a permanent-transitory income process, which is a common component of modern heterogeneous agents models.
Most recent draft - here
Instructional CEM example using the Rosenbrock function - here
Template for CEM application (change using your objective function) - here