Mehdi Es-hagh - Azad University branch of shahre-Rey Tehran. Royal institute of Technology Stockholm Sweden

In an adjustment process the observational vector may be included by different types of observations, I this situation it is suitable to estimate an unbiased estimate of veriance factor for each group of observations separately. The estimated posteriori variance factors for each group of observations is called variance component. The variance- covariance matrix of observations can be properly scaled by these variance components and the adjustment procedure itrated until the values of the variance components are converged to one. In this article the method of Best Quadratic Unbiased Estimation ( BQUS) and Best Quadratic Unbiased Non-negative Estimation (BQUNE) of variance components will be investigated in a simple network. In some cases these components may come out negative, in such cases the BQUNE method has tto be employed.

Adjustment , Variance Components , Weight