uniolympiads Ελληνικά

IMC 2004, Day 2, Problem 4

For n≥1n \ge 1 let MM be an n×nn \times n complex matrix with distinct eigenvalues λ1,λ2,…,λk\lambda_1, \lambda_2, \dots, \lambda_k, with multiplicities m1,m2,…,mkm_1, m_2, \dots, m_k, respectively. Consider the linear operator LML_M defined by LM(X)=MX+XMTL_M(X) = MX + XM^T, for any complex n×nn \times n matrix XX. Find its eigenvalues and their multiplicities. (MTM^T denotes the transpose of MM; that is, if M=(mk,l)M = (m_{k,l}), then MT=(ml,k)M^T = (m_{l,k}).)

Original from the official PDF4. For n ≥ 1 let M be an n × n complex matrix with distinct eigenvalues λ 1 , λ 2 , . . . , λ k , with multiplicities m 1 , m 2 , . . . , m k , respectively. Consider the linear operator L M defined by L M (X) = M X + XM T , for any complex n × n matrix X. Find its eigenvalues and their multiplicities. (M T denotes the transpose of M ; that is, if M = (m k,l ), then M T = (m l,k ).) [20 points]

This transcription was compared symbol by symbol with the original, but not yet by a person.

Time on this problem 0:00

Work for at least 45 minutes before opening the solution.

Official solution

Solution. We first solve the problem for the special case when the eigenvalues of MM are distinct and all sums λr+λs\lambda_r + \lambda_s are different. Let λr\lambda_r and λs\lambda_s be two eigenvalues of MM and v⃗r\vec{v}_r, v⃗s\vec{v}_s eigenvectors associated to them, i.e. Mv⃗j=λv⃗jM\vec{v}_j = \lambda\vec{v}_j for j=r,sj = r, s. We have Mv⃗r(v⃗s)T+v⃗r(v⃗s)TMT=(Mv⃗r)(v⃗s)T+v⃗r(Mv⃗s)T=λrv⃗r(v⃗s)T+λsv⃗r(v⃗s)TM\vec{v}_r(\vec{v}_s)^T + \vec{v}_r(\vec{v}_s)^T M^T = (M\vec{v}_r)(\vec{v}_s)^T + \vec{v}_r\left(M\vec{v}_s\right)^T = \lambda_r\vec{v}_r(\vec{v}_s)^T + \lambda_s\vec{v}_r(\vec{v}_s)^T, so v⃗r(v⃗s)\vec{v}_r(\vec{v}_s) is an eigenmatrix of LML_M with the eigenvalue λr+λs\lambda_r + \lambda_s.

Notice that if λr≠λs\lambda_r \ne \lambda_s then vectors u⃗,w⃗\vec{u}, \vec{w} are linearly independent and matrices u⃗(w⃗)T\vec{u}(\vec{w})^T and w⃗(u⃗)T\vec{w}(\vec{u})^T are linearly independent, too. This implies that the eigenvalue λr+λs\lambda_r + \lambda_s is double if r≠sr \ne s.

The map LML_M maps n2n^2–dimensional linear space into itself, so it has at most n2n^2 eigenvalues. We already found n2n^2 eigenvalues, so there exists no more and the problem is solved for the special case.

In the general case, matrix MM is a limit of matrices M1,M2,…M_1, M_2, \ldots such that each of them belongs to the special case above. By the continuity of the eigenvalues we obtain that the eigenvalues of LML_M are

  • 2λr2\lambda_r with multiplicity mr2m_r^2 (r=1,…,k)(r = 1, \ldots, k);
  • λr+λs\lambda_r + \lambda_s with multiplicity 2mrms2m_rm_s (1≤r<s≤k)(1 \le r < s \le k).

(It can happen that the sums λr+λs\lambda_r + \lambda_s are not pairwise different; for those multiple values the multiplicities should be summed up.)

Transcribed from the official solutions, page 2. This transcription has not been checked a second time; if in doubt, look at the original.

Archive problem: the statement is the official text (transcribed, with the original above). Subject classification and difficulty were estimated automatically; hints and a worked solution are added over time.

How did it go?