11.4. SCHUR’S THEOREM 193

upper triangular if it is of the form ∗ · · · ∗

. . ....

0 ∗

meaning that all entries are zero below the main diagonal, consisting of those entries of theform Tii.

Theorem 11.4.4 Let A be a real or complex n×n matrix. Then there exists a unitary matrixU such that

U∗AU = T, (11.3)

where T is an upper triangular matrix. If A has all real entries and eigenvalues, then Ucan be chosen to be orthogonal.

Proof: The theorem is clearly true if A is a 1× 1 matrix. Just let U = 1 the 1× 1matrix which has 1 down the main diagonal and zeros elsewhere. Suppose it is true for(n−1)× (n−1) matrices and let A be an n× n matrix. Then let v1 be a unit eigenvectorfor A. That is, there exists λ 1 such that

Av1 = λ 1v1, |v1|= 1.

By Theorem 11.4.3 there exists {v1, · · · ,vn}, an orthonormal set in Cn. Let U0 be a matrixwhose ith column is vi. Then from the above, it follows U0 is unitary. Then from the wayyou multiply matrices U∗0 AU0 is of the form

v∗1v∗2...v∗n

(

λ 1v1 Av2 · · · Avn

)=

λ 1 ∗ · · · ∗0... A1

0

where A1 is an n− 1× n− 1 matrix. Now by induction there exists an (n−1)× (n−1)unitary matrix Ũ1 such that

Ũ∗1 A1Ũ1 = Tn−1,

an upper triangular matrix. Consider

U1 ≡

(1 0

0 Ũ1

)From the way we multiply matrices, this is a unitary matrix and

U∗1 U∗0 AU0U1 =

(1 0

0 Ũ∗1

)(λ 1 ∗0 A1

)(1 0

0 Ũ1

)=

(λ 1 ∗0 Tn−1

)≡ T

11.4. SCHUR’S THEOREM 193upper triangular if it is of the form0 *meaning that all entries are zero below the main diagonal, consisting of those entries of theform T;;.Theorem 11.4.4 Let A be a real or complex n x n matrix. Then there exists a unitary matrixU such thatU*AU =T, (11.3)where T is an upper triangular matrix. If A has all real entries and eigenvalues, then Ucan be chosen to be orthogonal.Proof: The theorem is clearly true if A is a 1 x 1 matrix. Just let U = 1 the 1 x 1matrix which has 1 down the main diagonal and zeros elsewhere. Suppose it is true for(n—1) x (n—1) matrices and let A be an n x n matrix. Then let v; be a unit eigenvectorfor A. That is, there exists 2, such thatAv, =Aiu, |vi|=1.By Theorem 11.4.3 there exists {v1,--- ,v,}, an orthonormal set in C”. Let Up be a matrixwhose i” column is v;. Then from the above, it follows Up is unitary. Then from the wayyou multiply matrices UyAUp is of the formviv}. ( Aiw Av2 tee Av, ) =VyAy *K eee *K0: A\0where A, is an n—1 xn—1 matrix. Now by induction there exists an (n—1) x (n—1)unitary matrix U; such that aeUSAW = Th-1,an upper triangular matrix. Consider1 OU1 = ~0 U,From the way we multiply matrices, this is a unitary matrix and1 0 a 1 0 aU*UXAUgU = ~ rr * ~ j=( “7 *0 U; 0 Al 0 U, 0 Th- 1Il~