Estimating the Parameters of the Truncated Regression Model Using the Two Algorithms PSO and Quasi-Newton

Abstract

In this research estimation of the parameters of the truncated regression model first solution, which is known as particle swarm optimization algorithm, and the second is one of the conventional optimization algorithms; which is known as Quasi-Newton algorithm namely BFGS algorithm to reach the optimum values for these parameters. This research also proposes a hybrid algorithm, linking BFGS algorithm with PSO algorithm. To find the optimal values for these parameters, we are programming these algorithms using the ready matlab7.11(R2010b). Results show that the number of iterations resulting from the use of the hybrid algorithm (BFGS-PSO) is less than the number of iterations of the algorithm (BFGS) and that the results were obtained using the program Stata11 by the same amount of allowable errors.