Parameter estimation of nonlinear econometric models using particle swarm optimization

Global optimization is an essential component of econometric modeling. Optimization in econometrics is often difficult due to irregular cost functions characterized by multiple local optima. The goal of this paper is to apply a relatively new stochastic global technique, particle swarm optimization, to the well-known but difficult disequilibrium problem. Because of its co-operative nature and balance of local and global search, particle swarm is successful in optimizing the disequilibrium maximum likelihood function, providing better values than those reported in the literature obtained using other stochastic techniques. These encouraging results suggest that particle swarm optimization may be successfully applied to difficult econometrics problems, possibly in conjunction with existing methods.