The nearest generalized doubly stochastic matrix to a real matrix with the same first and second moments (Q1001132)

Let \(B\) be an \(n\times n\) real matrix. The authors explicitly give the nearest matrix (with respect to the Frobenius norm) \(A\) to \(B\), where \(A\) is a real generalized doubly stochastic matrix (a real matrix with the property that the sum of each column as well as of each row equals \(1\)) and \(e_1^TAe_1=e_1^TBe_1\), where \(e_1=(1,0,\dots,0)^T\). They also explicitly find the nearest generalized doubly stochastic matrix \(A\) to \(B\) with \(A\) satisfying \(e_1^TAe_1=e_1^TBe_1\) and \(e_1^TA^2e_1=e_1^TB^2e_1\), when such an \(A\) exists.