2. Preliminaries
In this section, we give some definitions and basic results for our main results.
A semiring [
1] is a set
with addition(+) and multiplication (·) such that
is a commutative semigroup with identity 0 and
is a semigroup with identity 1. Moreover, the operation · is distributive over +. A semiring is called anti-negative if only 0 has an additive inverse. A semiring is called commutative if
for all
.
The following are interesting examples of some anti-negative commutative semirings.
For a fixed positive integer
h, let
be the Boolean algebra [
6] of subsets of an
h-element set
and
denote the singleton subsets of
. Union is denoted by + (addition) and intersection by juxtaposition (multiplication); 0 denotes the empty set and 1 the whole set
. Under these two operations,
is a commutative anti-negative semiring, which is called generalized Boolean algebra. Then all of its elements, except 0 and 1, are zero-divisors because each proper subset of
has empty intersection with its complement. In particular, if
,
is called the binary Boolean algebra [
6].
Let
be any set with at least two elements. If
is totally ordered by < (i.e.,
or
for any two distinct elements
in
), then define
as
and
as
for all
. If
has a universal upper bound and a universal lower bound, then
is a semiring, which is called a chain semiring [
6].
Let
be the field of real numbers, let
0
1} be a subset of
. Define
and
for any
. Thus
is a chain semiring, which is called a fuzzy semiring [
7]. In particular, if we take
, a singleton set and we denote ∅ by 0 and
by 1, then this chain semiring is the binary Boolean algebra
, which is a subsemiring of any chain semiring. Since a Boolean algebra
(
) is not totally ordered under inclusion, it does not become a chain semiring.
In the following, will denote an arbitrary commutative anti-negative semiring. For all a, b , we supress the dot of a · b, and simply write ab. Let () be the set of all ( respectively) matrices with entries in . The addition and multiplication on are defined as usual.
In the following, we assume that integers p, q, m and n are positive integers with and .
Let be the (0,1)-matrix whose only th entry is 1, and be called a cell. O is the zero matrix, is the identity matrix and J is the matrix all of whose entries are 1.
A matrix is the hth full row matrix and is the kth full column matrix. A matrix L is called a full line matrix if or for some or for some .
A line of matrix is a column or a row of U.
A matrix has TR t (or term rank t) if the minimum number of lines needed to cover all nonzero entries of U is equal to t. We denote if U has TR t.
For matrices U, V , it is said that dominates (denoted or ) if implies for all h and k.
Lemma 1. For matrices U, V , we have
(1) ;
(2) ;
(3) implies that .
Proof. (1) It is trivial from the definition.
(2) If , then we can find t lines that cover all nonzero entries of U. If these lines cover all nonzero entries of V, then . However, If these lines cannot cover all nonzero entries of V, then . Thus, .
(3) If , then we can find t lines that cover all nonzero entries of V. Since , these lines can cover all nonzero entries of U. Thus . □
For a matrix U and lists and of column and row indices, respectively, denotes the submatrix constructed by deleting the columns and rows from U and denotes the submatrix constructed by taking the columns and rows from U.
For matrices , the matrix denotes the Schur or Hadamard product of U and V. That is, the entry of is .
A map is said to be linear if for all and .
If is a map, then T is called a -block map if there is a permutation matrix P and a permutation matrix Q, and with nonzero , such that
and , and for any or
and , and for any .
It is obvious that -block map is a linear map.
3. Linear Maps that Preserve TR of Matrices over Anti-Negative Commutative Semirings
In this section, we characterize the linear maps that preserve TR of matrices over anti-negative commutative semirings, which are contained in Theorem 2.
For a linear map , we say that T
- (i)
preserves TR t if whenever for all ;
- (ii)
doubly (or strongly) preserves TR t if if and only if for all ;
- (iii)
preserves TR if it preserves any TR t with .
Throughout this section, T denotes a linear map .
Lemma 2. Let T be a -block map. Then T doubly preserves any TR t for .
Suppose that T is a -block map, and U with with . Then or .
Consider the first case: . Since all entries of B are not zero, , and . Since the permuting columns and rows does not change the TR, we have
.
Thus T preserves TR t.
Conversely if , then
.
Thus . That is, T doubly preserves TR t.
Consider the second case: . As in the first case, a similar argument shows the same results. That is, T doubly preserves TR t.
Theorem 1. Let T doubly preserve TR 1. Then T is a -block map, and vice versa. (Here, we have either and , or and .)
Suppose that T doubly preserves TR 1. Then, the image of each line under T in is a line in since T preserves TR 1. Thus we assume that either or .
Consider the first case: . Assume that . Then, since is in both and and since and lies in the first row , h must be 1. However, then, for . Hence, has TR 1. However, , a contradiction. Thus we conclude that the image of any full column matrix is contained in a column matrix. By a similar argument, the image of any full row matrix is contained in a row matrix. And, since two columns have TR 2, the image of distinct full columns must be contained in distinct columns. Let be defined by if and define by if . Then, and are injective maps, and hence, and . Let and be a bijective maps such that and . Let and be the permutation matrices of order p and q, respectively, that correspond to the bijective maps and .
Thus we obtain that
and
, and there is some nonzero
such that
,
for every cell
. Therefore,
for every
. Thus
T is a
-block map.
Consider the second case:
. As in the first case, a similar argument implies that
and
. We obtain
for all
. Hence
for every
. This implies that
T is a
-block map.
Conversely, if T is a -block map, then T doubly preserves TR 1 by Lemma 2.
Lemma 3. Suppose that T preserves TR 1 and TR . Then we have
(1) T doubly preserves TR 1;
(2) T is a -block map.
(1) Consider the first case that : If has TR 1, then also has TR 1.
Conversely, if is the matrix with , and , then . However, since implies by assumption that . Thus . Let such that and with . Then . Thus by Lemma 1, which inequality is impossible. That is, T doubly preserves TR 1.
Consider the second case that : Assume that a TR 2 matrix is mapped to a TR 1 matrix. Then we may consider without loss of generality. Then, since T preserves TR 1 and TR t,
,
which is impossible. Hence, T doubly preserves TR 1.
(2) By (1), T doubly preserves TR 1. Hence T is a -block map by Theorem 1.
Lemma 4. Suppose that T preserves TR t.
(1) If and T does not preserve TR 1, then there is a matrix U such that and .
(2) If for some , then .
(1) Assume that T does not preserve TR 1 and for all U with . Then, there is a cell with . We may assume that without loss of generality. Since and T preserves TR t, we get . If we make , then we can take some cells such that for all , and . Since , there is a cell in whose image dominates two cells in , which contradicts for all U with in the assumption. This contradiction implies that there is a matrix U such that and .
(2) If , then since T preserves TR t. Assume that , and . Then there exists a matrix V such that and hence . However, by Lemma 1,
,
which is impossible. Therefore it follows that
Lemma 5. If T preserves any TR , but does not preserve TR 1, then , where J is the matrix with all entries 1.
By Lemma 4, if T does not preserve TR 1, then there is some matrix U such that and . So without loss of generality we may assume that .
Assume that . Then, . So we may assume that without loss of generality. Thus, there are cells, such that . Then, . However, and , which contradicts Lemma 4 (2). Thus, .
Lemma 6. Let . If and , then there is some , such that .
Assume that and . Then there are cells such that and . If then some cell must be a cell where , which is in contradiction with the assumption . Thus does not hold. That is, there is some h, , such that
.
Lemma 7. For , if T preserves consecutive TR t and TR , then it preserves TR 1.
If , we have finished. Assume that , and that T does not preserve TR 1. Then we have by Lemma 5. Since T preserves TR , .
Thus,
for
or
. Now, we may assume that for some
with
,
. This implies that
Now, without loss of generality, we may assume that there are
cells
such that
for
. Assume that the image of one cell in
dominates more than one cell in
. Without loss of generality, we may assume that
. Then,
, which is a contradiction since
, and hence
by Lemma 4 (2), but
Therefore, for each
,
dominates only one
. So, by permuting we may assume that
. Consider
. This matrix
Z must have TR
and dominates
. Applying Equation (
1) to Lemma 6, we can choose a cell
in
such that
. However,
, which is a contradiction.
This contradiction implies that T preserve TR 1.
Lemma 8. If T preserves TR t and TR s with , then it preserves TR 1.
Assume that T does not preserve TR 1. Then by Lemma 5. For any , we have by Lemma 1 and assumption. However, if we take , then must have TR s by assumption, while , a contradiction. That is, for arbitrary TR 1 matrix U. Therefore T preserves TR 1.
Lemma 9. If T preserves TR t and TR , then it preserves TR .
Let .
First, consider the case that and . Let be matrices of TR 1 such that . Since for some , we may assume that , for every . However, then while , a contradiction. Thus if , .
Second, consider the cases and . We may assume without loss of generality that and . Then there are r elements in whose sum dominates . Say, without loss of generality, that . Now, so that . However, since , it follows that and there are r elements of whose sum with has TR , say . Since , and . By the above case, we have a contradiction.
Thus T preserves TR .
Lemma 10. If T doubly preserves any one TR , then T preserves TR and hence T preserves TR 1.
Consider the first case : Assume that has TR 1. Then we may choose a matrix such that and . Since since T doubly preserves TR 2. Since , it follows that . Thus the Lemma holds in the case .
Consider the second case
: Let
and
. Assume that
. Say we may assume that
. Since
, it follows that
. So we may assume without loss of generality, that
and that
for some
. Thus, there are
s cells
in
such that
. Then
. Thus
. However,
which contradicts the assumption that
T doubly preserves TR
t. Hence
. Further,
by Lemma 4 (2), since
T doubly preserves TR
t. Therefore
, which implies that
T preserves TR
.
Moreover, T preserves TR 1 by Lemma 7.
Lemma 11. (1) If T preserves any two TR t and TR , then T is a -block map. (2) If T doubly preserves any one TR t, then we have that T is also a -block map.
(1) First case : Then T preserves TR 1 by Lemma 7.
Second case : Lemma 9 implies that T preserves TR . Hence T preserves TR 1 by Lemma 7.
Third case : Then by Lemma 8, T preserves TR 1.
Consequently, T preserves TR 1 by the above three cases. Hence, T doubly preserves TR 1 by Lemma 3. By Theorem 1, T is a -block map.
(2) By Lemma 10, T preserves TR 1. By Lemma 3, T doubly preserves TR 1. Thus T becomes a -block map by Theorem 1.
Now we have the main theorem:
Theorem 2. The following are equivalent for
T preserves any two TR t and TR s, with and
T doubly preserves any one TR t, with
T preserves TR;
T is the -block map.
It holds trivally that 3 implies 1 and 3 implies 2. Moreover, by Lemma 2, we have that 4 implies the other items 1, 2 and 3.
To show that 1 implies 4, suppose that T preserves TR t and TR s, with Then, by Lemma 11 (1), T is the -block map.
To show that 2 implies 4, suppose that T doubly preserves TR t. Then T is the -block map Lemma 11 (2).
Thus we obtained characterizations of the linear maps that preserve any two term rank between different matrix spaces.