GLOBAL OPTIMIZATION OF STATISTICAL FUNCTIONS WITH SIMULATED ANNEALING

Many statistical methods rely on numerical optimization to estimate a model's parameters. Unfortunately, conventional algorithms sometimes f ail. Even when they do converge, there is no assurance that they have found the global, rather than a local, optimum. We test a new optimiza tion algorithm, simulated annealing, on four econometric problems and compare it to three common conventional algorithms. Not only can simul ated annealing find the global optimum, it is also less likely to fail on difficult functions because it is a very robust algorithm. The pro mise of simulated annealing is demonstrated on the four econometric pr oblems.