E. Andruchow, L. E. MataLorenzo, A. Mendoza, L. Recht and A. Varela
Dedicated to the memory of Mischa Cotlar. Teacher and friend.
Abstract. We survey some results concerning the problem of finding the complex hermitian matrix or matrices of least supremum norm with variable diagonal. Some cualitative general results are given and more specific descriptions are shown for the 3 × 3 case. We also comment some results and examples concerning this approximation problem.
2000 Mathematics Subject Classification: 15A57, 47A58, 47C15
Key words and phrases. short curves, flag manifolds, matrix nearness, nearest diagonal.
Contents
1. Notation and Preliminaries
2. Previous results
3. Description of minimal hermitian matrices of 3 × 3 ]]>
4. Minimal matrices in a class
5. Different cases
6. The topology of the set of minimal matrices of 3 × 3
7. An example in 4 × 4 using minimal matrices
References
The problem of finding a complex hermitian matrix of least supremum norm with variable diagonal originally aroused when describing short curves in certain flag manifolds. A finite dimensional flag manifold is one such that its elements are chains of vector spaces included strictly. That is
The flags could be seen from different points of view. For example if we choose bases for the respective subspaces (that can be taken orthonormal), that is if
They can also be described choosing sets of orthonormal projections of rank one which sum gives the identity, that is:
Yet another way of representing flag manifolds is by the quotient between the group of unitary complex matrices over the subgroup of the unitary diagonals :
We will use the following notation
We will consider the quotient spaces

where is the usual operator (supremum) norm (respectively and for the antihermitian case).
The quotient space can be identified with the tangent of the space in the base point , since is an homogeneous space under the natural action of left multiplication of elements of :
Moreover, the Finsler metric considered in coincides with the quotient norm (1) of .
This will let us link geometric results to the problem we are interested to describe in this article. For instance, in theorem 1 of the next section, the existence of matrices reaching the minimum of the norm quotient (1) is related with the existence of curves in of the shortest posible length. In section 7, an example in of infinite different curves of the shortest length joining arbitrarily close points is shown.
]]> Operator approximation problems consist of finding, for a given operator, the element in some special class nearest to it, when distance is measured with a norm. These problems have been treated in the case of hermitian, positive and unitary approximants using different norms in [5], [6], [7], and others. The survey article [8] is related to matrix nearness. There, explicit formulas of operator approximation solutions are presented. Uniqueness results and algorithms for computing or estimating the minimal norm attained are also described, as well as the matrix or matrices sought in different contexts. Nevertheless, in that paper, the operator or supremum norm is not considered.The problem of finding the minimum of for a given matrix among all the diagonal matrices , and finding the matrix or matrices that realize the minimum, is indeed an operator approximation problem. It has a trivial translation to the problem of finding a real diagonal matrix that satisfies that and that is minimum.
In the case, some bounds of this minimum were obtained in [3]. In that work the calculation of this minimum is related to the estimation of bounds of the norm of the operator that for any matrix replaces all its diagonal entries by zeroes.
Definition 1. We will call a matrix minimal hermitian if
Remark 1. If is a minimal hermitian matrix and
 is a minimal antihermitian matrix.
 are eigenvalues of . ]]> The diameter of the spectrum of is .
 If is of then with .
In Theorem I of [4], Durán, MataLorenzo and Recht proved a result that, in the context of matrices, can be stated as follows:
Let and . If there exists a antihermitian minimal matrix that projects in , that is, and , then the curve
Remark 2. Note that this theorem implies that, in order to find the curve of minimum length such that , we have to find the matrix with the same offdiagonal entries as and with a diagonal that makes a minimal hermitian matrix.
This result shows the importance of the set of minimal matrices in the study of the shortest curves (in the geometric sense) in these homogeneus spaces.
]]> With the ideas of Section 5 of [4], resctricted to the context of matrices, the following characterization of minimal hermitian matrices can be obtained:Theorem 2. A hermitian matrix is minimal, if and only if, there exists a positive matrix such that,
 , where .
 The diagonal elements of the product are all zero.
3. Description of minimal hermitian matrices of 3 × 3
The previous theorem about minimal matrices together with auxiliary results in (see [1]), allow the following characterization:
Theorem 3. Let with . Then is minimal, if and only if, there exist two unitary eigenvectors,
 for the eigenvalue , and
 for the eigenvalue , ]]>
 ,
 ,
 with ,
 ,
 with , ,
 , , with:
 The eigenvector of has a zero coordinate and is of extremal type.
 The eigenvector of has no zero coordinates and is of non extremal type. ]]>
 When . In this case the surface is regular (a smooth manifold) and the method of Lagrange multipliers can be used to find the unique minimum in the class. In the figure to the right, the middle portion represents the component satisfying , and the dark point indicates the minimum.
 When and . In this case the surface is not regular, has one singular point which is the unique minimum in the class. In the figure to the right, two components of touch at the singular point which is the minimum.
 When exactly one coordinate of is null. In this case the surface is regular; the class has a unique minimum at the origin, . Observe that in the figure the vertical axis lies in .
 When exactly two coordinates of are null. In this case the surface is not regular along two curves, the two branches of the hyperbola shown in the figure, and the class has multiple minima, represented by the segment shown in the figure joining the two branches of the hyperbola.
 The family of curves will vary continuously with the parameter .
 Each of the curves of the family will parameterize the smaller arc of some circle in that joins to ; the arcs will not be great circles but for .
 The curve runs over a great circle in .
 The curve varies continuously with the parameter .
 The curve has constant speed in .
 The curve has constant speed in .
such that their respective coordinates have the same absolute value
Remark 3. Under these hypotheses, it can be proved that the eigenvectors and are triangular, that is, , and can represent the sides of a triangle (the sum of any two of them is always greater than the other one, see figure 1).
Parametrization of the set of minimal matrices of 3 × 3
Definition 2. Let . We will say that is of extremal type if there exist
Definition 3. Let . We will say that is of non extremal type if there exist:
such that
Figure 1:  Construction of and appearing in the parametrization. 
Remark 4. For matrices of both types (extremals and non extremals), the parameters and give the norm of , , and the trace of , .
In the previous Theorem 3 we have seen that , with is minimal, if and only if, there exist unitary eigenvalues and of and such that their coordinates have equal absolute value.
The following theorem gives a description of all the minimal hermitian matrices of in terms of the parameters , , , , , and .
Theorem 4. [Parametrization theorem] Let , then, is minimal, if and only if, verifies any of the following mutually exclusive cases:
4. Minimal matrices in a class
Every matrix can be writen in the form,
where and .
We already observed that, if is minimal hermitian (not null), the eigenvalues of are , and (with ).
If we consider the characteristic polynomial of :
Then for specific class of a matrix (that is for , and fixed)
]]> we consider the real manifold of given by the equation:Every minimal matrix of must belong to this manifold and must minimize
To simplify the expression of the map , the following linear change of variables can be introduced,
If the matrix is real with zero diagonal and we suppose that there is a diagonal that reaches the minimum of and such that has all of its eigenvalues of equal absolute value (that is, the spectrum of is and or has double multiplicity), then a precise formula of can be found in terms of the entries of (this proposition is proved in [10] motivated by results of [9]). We state it here as the following remark:
Remark 2. Let , with , be a matrix in and be a diagonal such that is minimum and the eigenvalues of have equal absolute value. Then must be of the form
6. The topology of the set of minimal matrices of 3 × 3
Using the theorem of parametrization and considering the sets
The following is a representation of the set of minimal hermitian matrices through .
7. An example in 4 × 4 using minimal matrices
We have remarked that in , despite of the existence of matrices that allow infinite minimizing diagonals, there are no multiplicity of short curves. This is not the case in .
Let us consider the manifold
Using a suitable description (see [2] and [11] for details) we can consider , where is the unit sphere in . This gives a better geometrical view of the example. Let us outline which are the properties of the curves mentioned at the beginning of the section.
Let be the point whose coordinates are both the North Pole, . Let be any point such that has higher latitude than in ( is closer to than ).
is going to be fixed so that is above the equator line (and is even higher). In [2], using the characterizations of minimal matrices seen previously, a family of minimal curves , was constructed, all joining to , with the following properties.

The following is a representation of these curves in :
[1] Andruchow, Esteban; MataLorenzo, Luis E.; Recht, Lázaro; Mendoza, Alberto; Varela, Alejandro. Minimal matrices and the corresponding minimal curves on flag manifolds in low dimension, preprint. [ Links ]
[2] Andruchow, Esteban; MataLorenzo, Luis E.; Recht, Lázaro; Mendoza, Alberto; Varela, Alejandro. Infinitely many minimal curves joining arbitrarily close points in a homogeneous space of the unitary group of a algebra. Rev. Un. Mat. Argentina 46 (2005), no. 2, 113–120 (2006). [ Links ]
[3] Bhatia, R., Choi, M.D., Davis, C. Comparing a matrix to its offdiagonal part. The Gohberg anniversary collection, Vol. I (Calgary, AB, 1988), 151–164, Oper. Theory Adv. Appl., 40, Birkhauser, Basel, 1989. [ Links ]
[4] Durán, C.E., MataLorenzo, L.E., Recht, L., Metric geometry in homogeneous spaces of the unitary group of a Calgebra: Part I–minimal curves, Adv. Math. 184 No. 2 (2004), 342366. [ Links ]
[5] Fan, K., Hoffman, A.J., Some metric inequalities in the space of matrices, Proc. Amer. Math. Soc., 6 (1955) 11116. [ Links ]
[6] Halmos, P.R., Positive approximants of operators, Indiana Univ. Math. J., 21 (1972), 951–960. [ Links ]
[7] Halmos, P.R., Spectral approximants of normal operators, Proc. Edinburgh Math. Soc., 19 (1974), 51–58. [ Links ]
[8] Higham, N.J., Matrix nearness problems and applications, Applications of Matrix Theory, Oxford University Press, 1989, 1–27. [ Links ]
[9] MataLorenzo, L.E., Recht, L. Convexity properties of , Linear Algebra and its Applications 315 (2000), 2538. [ Links ]
[10] Varela, A., Villagra, M.P., Algunos resultados de minimización de matrices de diagonal variable, preprint. [ Links ]
[11] Whittaker, E. T. A Treatise on the Analytical Dynamics of Particles and Rigid Bodies, Cambridge University Press, London, 1988. [ Links ]
Esteban Andruchow
Instituto de Ciencias,
Universidad Nacional de Gral. Sarmiento,
J. M. Gutierrez 1150, (1613) Los Polvorines,
Argentina
eandruch@ungs.edu.ar
Luis E. MataLorenzo ]]>
Universidad Simón Bolívar,
Apartado 89000, Caracas 1080A, Venezuela
lmata@usb.ve
Alberto Mendoza
Universidad Simón Bolívar,
Apartado 89000, Caracas 1080A, Venezuela
jacob@usb.ve
Lázaro Recht
Universidad Simón Bolívar,
Apartado 89000, Caracas 1080A, Venezuela ]]>
recht@usb.ve
Alejandro Varela
Instituto de Ciencias,
Universidad Nacional de Gral. Sarmiento,
J. M. Gutierrez 1150, (1613) Los Polvorines,
Argentina
avarela@ungs.edu.ar
Recibido: 10 de abril de 2008
Aceptado: 15 de mayo de 2008