Adaptive Markov chain Monte Carlo sampling and estimation in Mata

I describe algorithms for drawing from distributions using adaptive Markov chain Monte Carlo (MCMC) methods, introduce a Mata function for per- forming adaptive MCMC, amcmc(), and a suite of functions amcmc_*() allowing an implementation of adaptive MCMC using a structure. To ease use in application to estimation problems, amcmc() and amcmc_*() can be used in conjunction with models set up to work with Mata’s moptimize( ) or optimize( ), or with stand-alone functions. I apply the routines in a simple estimation problem, in drawing from a distributions without a normalizing constant, and in Bayesian estimation of a mixed logit model.