G-stochastic matrices and linear preservers of G-majorization

Let Mn be the algebra of all n by n real or complex matrices. A nonneg- ative matrix R in Mn which all it's row sums are equal one is said to be row stochastic matrix. A column stochastic matrix is the transpose of a row stochastic matrix. A matrix D in Mn with the property that D and D^t are row stochastic matrices is said to be doubly stochastic matrix. A matrix R in Mn which all it's row sums are equal one is said to be g-row stochastic matrix. A matrix C in Mn which all it's column sums are equal one is said to be g-column stochastic matrix. A matrix D in Mn with the property that D and D^t are g-row stochastic matrices is said to be g-doubly stochastic matrix. The matrix B is said to be gw-majorized (or gs-majorized) by A if there exists an n by n g-row (or g-doubly) stochastic matrix R such that B=RA, and denoted by AgwB(orA gs B). we will characterize all linear operators that : (1) preserve (or strongly preserve) gw-majorization on Rn and Mn. (2) preserve (or strongly preserve) gs-majorization on Mn,m.