Computation of Parameters Using Genetic Algorithm and Sequential Quadratic Programming in Sampling

The sampling literature contains many examples of estimators of population parameter. To deal with this problem many authors have suggested family of estimators of population parameter. But in the case of generalization of these estimators, estimation of optimum values is a problem. Some authors can define estimator replacing the unknown parameters by their sample estimates. To get the optimum estimator, one need to solve complex mean square error equation with many parameters and nonlinear constraints. In this study we have tried to get these optimum parameter in stratified random sampling using genetic algorithms and sequential quadratic programming. A numerical example is also done to compare these algorithms. The results show that genetic algorithm is more efficient than sequential quadratic programming to solve the complex model with more parameters under non-linear constraints.