Vous êtes sur la page 1sur 13

The sum of matrices of the same kind

Sum of matrices
To add two matrices of the same kind, we simply add the corresponding elements.

Sum properties
Consider the set S of all n x m matrices (n and m fixed) and A and B are in S. From the properties of real numbers it's immediate that

A + B is in S the addition of matrices is associative in S A+0=A=0+A with each A corresponds an opposite matrix -A A+B=B+A

Scalar multiplication
Definition
To multiply a matrix with a real number, we multiply each element with this number.

Properties
Consider the set S of all n x m matrices (n and m fixed). A and B are in S; r and s are real numbers. It is not difficult to see that:
r(A+B) = rA+rB (r+s)A = rA+sA (rs)A = r(sA) (rA)T = r. AT

Sums in math
Because in the following, there is an intensive use of the properties of sums, the reader who is not familiar with these properties must read first Sums in math . Remark. In this html document, for convenience, we'll write the word sum instead of the sigma sign.

Multiplication of a row matrix by a column matrix

This multiplication is only possible if the row matrix and the column matrix have the same number of elements. The result is a ordinary number ( 1 x 1 matrix). To multiply the row by the column, you have to multiply all the corresponding elements, then make the sum of the results. Example.
[1] [2 1 3]. [2] = [19] [5]

Multiplication of two matrices A.B


This product is defined only if A is a (l x m) matrix and B is a (m x n) matrix. So the number of columns of A has to be equal to the number of rows of B. The product C = A.B then is a (l x n) matrix. The element of the i-th row and the j-th column of the product is found by multiplying the ith row of A by the jth column of B.
ci,j = sumk (ai,k.bk,j)

Examples.
[1 2][1 3] = [5 7] [2 1][2 2] [4 8] [1 3][1 2] = [7 5] [2 2][2 1] [6 6] [1 1][2 [1 1][-2 2] = [0 0] -2] [0 0] [ 1, 16, 5 ] [ 8, 23, 18 ] [ 4, 8, 9 ]

[ 1, 3, 2 ] [ 3, -1, 4 ] [ 4, 5, 3 ] [ -2, 3, 1 ] = [ 2, 2, 1 ] [ 2, 4, -1 ]

From these examples we see that the product is not commutative and that there are zero divisors. Zero divisors are matrices different from a zero matrix, such that the product is a zero matrix. Application A matrix A is called idempotent if and only if A2 = A. Given:
[1 A = [0 [0 b 0 0 c] 2] 1]

Find the set of all 3 x 3 matrices of type A such that A is idempotent.

Solution: We calculate A2.


[1 [0 [0 <=> <=> b 0 0 2c+2b] 2 ] 1 ] A2 = A 2c + 2b = c c = -2b

All requested matrices are


[1 [0 [0 b 0 0 -2b] 2 ] 1 ] with b in R

Properties of multiplication of matrices


Associativity
If the multiplication is defined then A(B.C) = (A.B)C holds for all matrices A,B and C. Proof: We'll show that an element of A(B.C) is equal to the corresponding element of (A.B)C First we calculate the element of the ith row and jth column of A(B.C)
Let D denote B.C, then dk,j = sump bk,p.cp,j Let E denote A.D then ei,j = sumk ai,k.dk,j (2) (1) in (2) gives ei,j = sumk ai,k.(sump bk,p.cp,j) <=> ei,j = sumk,p ai,k.bk,p.cp,j (1)

So the element of the ith row and jth column of A(B.C) is sumk,p ai,k.bk,p.cp,j (3)

Now we calculate the element of the ith row and jth column of (A.B)C
Let D' denote A.B, then di,p' = sumk ai,k.bk,p Let E' denote D'C then ei,j' = sump di,p'.cp,j (4) (5)

(4) in (5) gives ei,j' = sump (sumk ai,k.bk,p).cp,j

<=>

ei,j' = sumk,p ai,k.bk,p.cp,j

So the element of the ith row and jth column of (A.B)C is sumk,p ai,k.bk,p.cp,j (6) From (3) and (6) => A(B.C) = (A.B)C

Distributivity
If the multiplication is defined then A(B+C) = A.B+A.C and (A+B).C = A.C+B.C hold for all matrices A,B and C. This theorem can be proved in the same way as above.

Theorem 1
For each A, there is always an identity matrix E and an identity matrix E' so that A.E = A and E'.A = A. If A is a square matrix then E = E'.

Theorem 2
(A.B)T = BT .AT

This theorem can be proved in the same way as above. Example : If we transpose the product
[ 2 4 ] [x] [ 3 8 ] [y]

we get
[x y ] [ 2 3 ] [ 4 8 ]

Theorem 3
If the multiplication is defined then we have for any A
A.0 = 0 = 0.A

Theorem 4
r and s are real numbers and A , B are matrices. If the multiplication is defined then (rA) (sB) = (rs)(AB) This theorem can be proved in the same way as above.

Theorem 5
If D = diag(a,b,c) then D.D = diag( a2 , b2 , c2)

D.D.D = diag( a3 , b3 , c3) .....

This property can be generalized for D = diag(a,b,c,d,e,...,l).

The transposed matrix of a matrix


The n x m matrix B is the transposed matrix of the m x n matrix A if and only if The ith row of A = the ith column of B for (i = 1,2,3,..m) So ai,j = bj,i
The transposed matrix [7 [0 [3 1 ] 5 ] 4 ] T = of A is denoted T(A) or AT [7 [1 0 5 3] 4]

Level 2 problems

A and B are n x n matrices. Investigate if (A + B)2 = A2 + 2.A.B + B2

(A + B)2

= (A + B).(A + B) = A.A + A.B + B.A + B.B = A2 + A.B + B.A + B2

Since the product is not commutative, we do NOT have (A + B)2 = A2 + 2.A.B + B2

We say that two matrices A and B commute, if and only if AB = BA. Show that all matrices of the form [a -b]

[b commute.

a]

[a [b and [a' [b'

-b] [a' a] [b'

-b'] a']

= [aa'-bb' [ba'+ab'

-ab'-ba'] -bb'+aa']

-b'] [a a'] [b

-b] a]

= [aa'-bb' [ba'+ab'

-ab'-ba'] -bb'+aa']

Prove that for each n x n matrix A ,

AT .A is symmetric

We'll prove that AT .A is equal to its transpose. (AT .A)T = AT .(AT)T = AT .A

Prove that if matrix A is skew-symmetric, then A.A is symmetric.

A is skew-symmetric <=> A = - AT . So, (A.A)T = AT .AT = (- A ).(- A ) = A.A Hence, A.A is symmetric.

Given

X = [x

y] and A is a 2 x 2 matrix .

All elements of the matrices are real numbers. Prove that X . AT . A . XT can't be negative.

First we see that X . AT . A . XT is a 1 x 1 matrix. It is a number, so maybe it can be negative. But, X . AT A . XT Then X . AT . A . XT = (X . AT ).(X . AT )T = [a = a2 So, X . AT . A . XT b] .[a + b2 negative. b]T = is a 1 x 2 matrix, say [a = (X . AT )T = [a b]T b].

can not be

Determine for which values of a and b the following system has no solutions.
/ ax + y + 2z = 0 | x + 2y + z = b \ 2x + y + az = 0

The matrix formed by the coefficients is


[ a [ 1 [ 2 1 2 1 2] 1] a]

and the determinant D= 2a2 - 2a - 4. D = 0 <=> a = 2 of a = -1 1. If D not 0, then the system has exactly 1 solution. 2. If a = 2, the matrix formed by the coefficients is
3. 4. [ 2 5. [ 1 6. [ 2 1 2 1 2] 1] 2]

The rank of this matrix is 2. We can choose the third equation as side equation and z as a side unknown. The characteristic determinant is here
| 2 | 1 | 2 1 2 1 0| b| = 0 0|

The system has always solutions when a = 2. 7. If a = -1, the matrix formed by the coefficients is
8. 9. [-1 1 10. [ 1 11. [ 2 2] 2 1] 1 -1]

The rank of this matrix is 2. We can choose the third equation as side equation and z as a side unknown. The characteristic determinant is now
|-1 | 1 | 2 1 2 1 0| b| = 3b 0|

If b not 0, the system has no solutions


Conclusion: the system has no solutions if and only if (a = -1 and b not 0)

Calculate the m-values such that the following system has more than one solution.
/ (m+2)x + 2y + 4z = 3m | -mx + 5y + 2mz = -2m \ 2x + 7y + 6z = 1

If the determinant of the matrix of the coefficients is not zero, then the system has a unique solution. So, this determinant must be zero. This is the case for m = 1 or m = -10/7. Case : m = 1 If we substitute 1 for m in the system, we immediately see that the third equation is the sum of the previous equations. The third equation is superfluous. We cancel

this equation. Now, the rank of the matrix of the coefficients is 2 and we can choose z as a side-unknown. The system has infinitely many solutions. Case : m = -10/7 we substitute -10/7 for m in the system.We simplify a bit.
/ 4x + 14y + 28z = -30 | 10x + 35y - 20z = 20 \ 2x + 7y + 6z = 1

The rank of the matrix of the coefficients is 2. We can't use x an y as main unknowns. The unknowns y and z can be used as main unknowns with the first and second equation as main equations. The system has a solution if and only if the characteristic determinant of the side equation is zero. In this system, the characteristic determinant is not zero. The system has no solutions if m = -10/7.

Given

[ 1 A = [ m [ m + 1

m - 1 2 m - 2 3 m - 3

2 m - 3 2 m.m - 1

] ] ]

Calculate the condition for m such that A is regular. Assume that m satisfies this condition and consider the system with A as coefficient matrix. / x + (m - 1)y + (2 m - 3)z | | m x + (2 m - 2)y + 2 z | \ (m + 1)x + (3 m - 3)y + (m.m - 1)z = 1 = 0 = 0

Now, let x = 1 and calculate the values of m such that the system has a solution for y and z.

Det(A) = m (1 - m) (m - 2)

2 m is NOT in { 0, 1, 2 }

A is regular if and only if

With x = 1, the system becomes / (m - 1)y | | (2 m - 2)y | \ (3 m - 3)y + (2 m - 3)z + 2 z = 0 = - m

+ (m.m - 1)z

= -(m + 1)

The coefficient matrix of the first and the second equation is

[m - 1 [2 m - 2

2 m - 3 2

] ]

The determinant is -4 (m - 1) (m - 2) and since m is not in { 0, 1, 2 }, the determinant is not 0. So, we consider the first and the second equation as main equations and the system has a solution for y and z if and only if the characteristic determinant of the third equation is 0.
The condition is |m - 1 2 m - 3 0 | |2 m - 2 2 - m | = 0 |3 m - 3 m.m - 1 - m - 1| <=> 2 (m + 4) (m - 1) (m - 2) = 0

Since m is not in { 0, 1, 2 }, there are no real values of m such that the system has a solution for y and z.

The matrix x is a 2 x 2 matrix. Calculate three solutions of the quadratic equation x2 - x = 0

x2 <=>

x = 0

x (x-I) = 0

x = 0-matrix and x = Identity matrix are two solutions, but we know that x (x-I) can be zero without x = 0 or x = I. To find another solution of the given equation we'll search for a matrix
[ 1 [ [ c = x b ] [ 1 b ] [ 1 b ] b ] ] d ]

x = x2 <=> [ 1

such that

x2

= x

[ [ c <=> [ 1 + b c [ [ c + d c [ b c [ [ c d

].[ d ] [ c b + b d c b + d2 ] b d

] = [ d ] [ c ] ] = [ 1 [ [ c [0 [ [0

] d ] b ] ] d ] 0 ] ] 0 ]

<=>

] ] =

b c + d2 - d ]

Since we need only one solution, we choose d = 0 and c = 0. Then b = an arbitrary number. So, even with d = 0 and c = 0, we find an infinity number of solutions.
[ 1 [ [ 0 b ] ] 0 ]

x =

Level 3 problems

Given

[-2 A = [ [ 1

-9 ] ] . Prove that 4 ]

[1-3n = [ [ n

-9n ] ] 1+3n ]

We'll prove this by complete induction. a) The property is trivial for n = 1 b) Assume the property is true for n = k. Then we have to prove that the property is true for n = k+1. k+1 A k [1-3k = A .A = [ [ k = [1-3(k+1) [ [ k+1 -9k ] [-2 ].[ 1+3k ] [ 1 -9(k+1) ] ] 1+3(k+1) ] -9 ] [-2-3k ] = [ 4 ] [k+1 -9-9k] ] 3k+4]

Given :

A =

[1 [0 [0

1 1 0

0 ] 0 ] 1 ]

Show that A is regular. n Calculate A . Calculate the real numbers a and b such that matrix) A2 + a A + b I = 0 ( I is the 3 x 3 identity

Show that there are real numbers A-n = c0.I + c1.A2 + c2.A3 + c3.A4

c0,c1,c2, ... ,cn such that + c4.A + ... + cn.An

Det(A) = 1 , so A is regular. 2 A = [1 [0 [0 2 1 0 0 ] 0 ] 1 ] 3 A = [1 [0 [0 3 1 0 0 ] 0 ] 1 ]

From this we'll prove by complete induction that n A = [1 [0 [0 n 1 0 0 ] 0 ] 1 ]

This expression is true for n = 1 Assume the property is true for n = k. Then we have to prove that the property is true for n = k+1. k+1 A k [1 = A .A = [0 [0 k 1 0 0 ][1 0 ][0 1 ][0 A2 <=> 0 ] 0 ] [1 [0 2 1 0 ] [1 1 1 0 0 ] [ 1, k + 1, 0 ] 0 ] = [ 0, 1, 0 ] 1 ] [ 0, 0, 1 ]

+ a A + b I = 0 1 1 0 ] [1 0 1 0 ] [0 0 0

0 ] + a [0

0 ] + b [0

0 ] = [0

0 ] <=>

[0

1 ]

[0

1 ]

[0

1 ]

[0

1 + a + b = 0 2 + a = 0 <=> a = -2, b = 1

From this we have A2 <=> 2 A - A2 = I <=> A (2 I - A ) = I From this, we see that (2 I - A ) is the inverse of A A-1 => A-n = 2 I - A = (2 I - A )n - 2 A + I = 0

=> A-n = a polynomial in A of degree n, with real coefficients => there are real numbers A-n = c0.I + c1.A c0,c1,c2, ... ,cn such that c3.A3 + c4.A4 + ... + cn.An

+ c2.A2 +

Vous aimerez peut-être aussi