General unbiased estimating equations for variance components in linear mixed models

This paper introduces a general framework for estimating variance components in the linear mixed models via unbiased equations, which include some well-used estimators such as restricted maximum likelihood estimator. We derive asymptotic covariance matrices and second-order biases under equations without assuming normality of underlying distributions identify class components. It is also shown that do not depend on whether regression coefficients are estimated by generalized or ordinary least squares methods. carry out numerical studies to check performance proposed methods based typical models.