Let V be the vector space of all n×n matrices over R. For i,j∈{1,2,...,n} denote by Eij the matrix with a unigue 1 at the intersection of the row i and the column j, and with the rest elements equal to 0: eij=1 and est=0 for all s=i and t=j. Then (Eij ∣ i,j∈{1,...,n}) is a basis of the vector space V. Indeed, if i=1∑nj=1∑naijEij=O where the matrix O contains of only zeros, then aij=0 for all i,j∈{1,2,...,n}. Consequently, the matrices Eij are linear independent. Also each matrix M with entries mij can be represented as M=i=1∑nj=1∑nmijEij, and therefore, (Eij ∣ i,j∈{1,...,n}) indeed a basis of V. It follows that dim(V)=n2.
Further, let Wn={An×n∈V ∣ An×n is upper triangular}. Let An×n,Bn×n be elements of Wn and α,β be elements of R. Since the elements under the main diagonal in matrices An×n,Bn×n are zeros, it follows that the elements under the main diagonal in the matrix αAn×n+βBn×n is zeros as well. Therefore, αAn×n+βBn×n is in Wn, and we conclude that Wn is a subspace of V.
Comments