Matrix Theory: Rank of Matrix

-
System Of linear equations

FY BTECH SEM-I
MODULE:2
Sub-module: 2.3
 A SYSTEM OF LINEAR EQUATIONS
• Consider a system of m linear vector of order 𝑚 × 1
equations in n unknowns, say 𝑿 = [𝒙𝟏 𝒙𝟐 𝒙𝟑 … 𝒙𝒏 ]𝑻 is the column
𝑥1 , 𝑥2 , 𝑥3 , … . 𝑥𝑛 vector of order (𝑛 × 1)
𝑎11 𝑥1 + 𝑎12 𝑥2 + ⋯ + 𝑎1𝑛 𝑥𝑛 = 𝑏1 • Any vector U satisfying AU = B is said to
𝑎21 𝑥1 + 𝑎22 𝑥2 + ⋯ + 𝑎2𝑛 𝑥𝑛 = 𝑏2 be a solution of AX = B.
𝑎31 𝑥1 + 𝑎32 𝑥2 + ⋯ + 𝑎3𝑛 𝑥𝑛 = 𝑏3 • The matrix [A,B] i.e., the matrix formed
………………………………………………..𝑎𝑚1 𝑥1 + by the coefficients and the constants is
𝑎𝑚2 𝑥2 + ⋯ + 𝑎𝑚𝑛 𝑥𝑛 = 𝑏𝑚 called the augmented matrix.
• This system can be written compactly • A system 𝐴𝑋 = 𝐵 is
in matrix notation as 𝑨𝑿 = 𝑩 (i) Homogeneous if 𝐵 = 𝑂 and
𝑨 = 𝒂𝒊𝒋 : matrix of coefficients (ii) Non – homogeneous if 𝐵 ≠ 𝑂
𝒎𝑿𝒏
𝑩 = 𝒃𝟏 𝒃𝟐 𝒃𝟑 … 𝒃𝒎 𝑻 is the column
 Different cases for solution
Consistent
Inconsistent
(possess at least one solution)
• (i) Unique Solution: • (ii) Infinite Solutions: • (iii) No Solution:
Ex: 4𝑥1 + 3𝑥2 = 11 , Ex: 4𝑥1 + 3𝑥2 = 11 , Ex: 4𝑥1 + 3𝑥2 = 11 ,
4𝑥1 − 3𝑥2 = 5 8𝑥1 + 6𝑥2 = 22 8𝑥1 + 6𝑥2 = 20
• Solving we get, has more solutions, say has no solution at all.
𝑥1 = 2 and 𝑥2 = 1 2,1 𝑇 or 0,11/3 𝑇 .
11−4𝑘 𝑇
[𝑘, ] is a general
3
solution for all k
 NON – HOMOGENEOUS SYSTEM
OF LINEAR EQUATIONS
(i) n EQUATIONS IN n UNKNOWN:
Method 1: by finding 𝑨−𝟏
Consider 𝐴𝑋 = 𝐵, where A is a non – singular 𝑛 × 𝑛 matrix, 𝑋 is 𝑛 × 1 vector and B is
𝑛 × 1 matrix then the system has unique solution.
• Steps:
1. Write 𝐴𝑋 = 𝐵
2. Check that 𝐴 ≠ 0
3. Find 𝐴−1 by any suitable method.
4. Solution is given by 𝑿 = 𝑨−𝟏 𝑩.
• Note: If A is singular matrix, then this inverse method fails. In that case the system may
have infinitely many solutions or none at all.
 NON – HOMOGENEOUS SYSTEM
OF LINEAR EQUATIONS
• (ii) m EQUATIONS IN n UNKNOWN:
• Method :2
• Case (i) : 𝑅𝑎𝑛𝑘 𝐴 < 𝑅𝑎𝑛𝑘 [𝐴, 𝐵]
In this case the equations are inconsistent i.e., they have no solution.
• Case (ii): 𝑅𝑎𝑛𝑘 𝐴 = 𝑅𝑎𝑛𝑘 [𝐴, 𝐵] = 𝑟
In this case the equations are consistent i.e., they possess a solution.
• Further,
(a) If 𝑟 = 𝑛 i.e if the rank of A is equal to the number of unknowns, the system has unique
solutions.
(Also note that the system has unique solution if the coefficient matrix is non –
singular).
(b) If 𝑟 < 𝑛, if the rank of A is less than the number of unknown the system has infinite
solutions.
• In this case 𝑛 – 𝑟 unknowns called parameters can be assigned arbitrary values.
• The remaining unknowns then can be expressed in terms of these parameters.
 NON – HOMOGENEOUS SYSTEM
OF LINEAR EQUATIONS
• Working Rule:
1. Write the given system in the matrix form
𝐴𝑋 = 𝐵. No System is
If Rank A =
inconsistent
2. Apply row transformations on A as well as on Rank [A, B]
(No Solution)
the column matrix B i.e. on the augmented
matrix [A, B] till you get an row echelon form. Yes
rankA=no.o
3. We know that the rank of a matrix in echelon System has unique
f variable
form is equal to the number of non-zero rows. r=n
Yes solution
Determine the rank of A and the rank of the No
augmented matrix [A, B] and find the solution by then r < n
using following cases.
System has
infinitely many
solutions
 Example 1
• Test the consistency of the equations • Applying 𝑅3 − 3𝑅2 , we get
1 1 1 3
• 𝑥 + 𝑦 + 𝑧 = 3, 𝑥 + 2𝑦 + 3𝑧 = 4, 𝑥 + 4𝑦 + 9𝑧 = 6. 𝐴: 𝐵 ~ 0 1 2 1
• Solution: The system of equations can be written 0 0 2 0
1 1 1 𝑥 3 • ∴ rank of [𝐴: 𝐵] = rank of 𝐴 = 3.
as 𝐴𝑋 = 𝐵, i.e. 1 2 3 𝑦 = 4
1 4 9 𝑧 6 • ∴ The system has unique solution.
• ∴ the augmented matrix can be written as • The reduced form of equations can be written as
1 1 1 3
𝐴: 𝐵 = 1 2 3 4 • 𝑥 + 𝑦 + 𝑧 = 3, 𝑦 + 2𝑧 = 1 , 2𝑧 = 0
1 4 9 6 • ∴ (iii) ⇒ 𝑧 = 0 substituting this value in (ii) 𝑦 = 1
• Applying 𝑅2 − 𝑅1 𝑎𝑛𝑑 𝑅3 − 𝑅1 we get • ∴ (i) ⇒ 𝑥 + 1 + 0 = 3 ⇒ 𝑥 = 2
1 1 1 3
𝐴: 𝐵 ~ 0 1 2 1 𝑥 2
0 3 8 3 • Hence the solution set can be written as 𝑦 = 1
𝑧 0
 Example 2
𝑥1 + 2𝑥2 + 𝑥3 = 2 • Applying 𝑅3 − 2𝑅2 , we get
• Solve 2𝑥1 + 4𝑥2 + 3𝑥3 = 3 1 2 1 2
3𝑥1 + 6𝑥2 + 5𝑥3 = 4 𝐴: 𝐵 ~ 0 0 1 −1
• The system of equations can be represented by the 0 0 0 0
1 2 1 𝑥1 2 • Hence 𝜌 𝐴 = 𝜌[𝐴: 𝐵], therefore the system is
matrix equation as, 2 4 3 𝑥2 = 3 consistent.
3 6 5 𝑥3 4 • Further rank 𝑟 = 2 < 3 (number of variables),
• Where the augmented matrix is therefore the system has infinite solutions.
1 2 1 2 • ∴ 𝑛 − 𝑟 = 3 − 2 = 1 (free variable)
• 𝐴: 𝐵 = 2 4 3 3
3 6 5 4 • The reduced form of the linear equations can be
written as, 𝑥1 + 2𝑥2 + 𝑥3 = 2 , 𝑥3 = −1
• Applying elementary row transformations, the
matrix 𝐴: 𝐵 can be reduced to Echelon form. • Let 𝑥2 = 𝑘, an arbitrary constant. ∴ 𝑥1 = 3 − 2𝑘
𝑥1 3 − 2𝑘
• Applying 𝑅2 − 2𝑅1 𝑎𝑛𝑑 𝑅3 − 3𝑅1 we get
1 2 1 2 • Hence 𝑥2 = 𝑘 infinite solutions as k varies.
𝑥3 −1
𝐴: 𝐵 ~ 0 0 1 −1
0 0 2 −2
 Example 3
• Are the following equations consistent? Justify. • Applying 𝑅12 , we get
2𝑥 + 𝑦 + 𝑧 = 4 1 1 1 2
𝑥+𝑦+𝑧=2 𝐴: 𝐵 ~ 2 1 1 4
5𝑥 + 3𝑦 + 3𝑧 = 6 5 3 3 6
• Solution: The system of linear equations can be • Applying 𝑅2 − 2𝑅1 𝑎𝑛𝑑 𝑅3 − 5𝑅1 , we get
written in the matrix form 𝐴𝑋 = 𝐵 i.e 1 1 1 2
2 1 1 𝑥 4 𝐴: 𝐵 ~ 0 −1 −1 0
1 1 1 𝑦 = 2 0 −2 −2 −4
5 3 3 𝑧 6 • Applying 𝑅3 − 2𝑅1 , we get
• ∴ the augmented matrix can be written as 1 1 1 2
2 1 1 4 𝐴: 𝐵 ~ 0 −1 −1 0
𝐴: 𝐵 = 1 1 1 2 0 0 0 −4
5 3 3 6 • ⇒ 𝑟𝑎𝑛𝑘 𝑜𝑓 𝐴 = 2 and rank of 𝐴: 𝐵 = 3.
• i.e 𝜌(𝐴) ≠ 𝜌 𝐴: 𝐵 .
• Hence the given system of linear equation is
inconsistent and therefore has no solution.
 Example 4
• Solve the system of equations. • Applying 𝑅3 − (𝑅1 + 𝑅2 ) ,we get
𝑥1 − 𝑥2 + 2𝑥3 + 𝑥4 = 2 1 −1 2 1 2
3𝑥1 + 2𝑥2 + 𝑥4 = 1 𝐴∶𝐵 ~ 3 2 0 1 1
4𝑥1 + 𝑥2 + 2𝑥3 + 2𝑥4 = 3 0 0 0 0 0
• Solution: The system of equations can be • Applying 𝑅2 − 3𝑅1 , we get
represented by the matrix equation as, 1 −1 2 1 2
𝐴 ∶ 𝐵 ~ 0 5 −6 −2 −5
𝑥1 0 0 0 0 0
1 −1 2 1 𝑥 2
2 • Hence 𝜌 𝐴 = 𝜌 𝐴: 𝐵 = 2, the system is
3 2 0 1 𝑥3 = 1
consistent
4 1 2 2 𝑥4 3
• Rank 𝑟 = 2 < 4 (Number of variables),
• where the augmented matrix is therefore the system has infinite solutions.
1 −1 2 1 2 • ∴ 𝑛 − 𝑟 = 4 − 2 = 2 (free variables)
• 𝐴∶𝐵 = 3 2 0 1 1
4 1 2 2 3
 Example 4 contd..
1
• The reduced form of the linear equations is 𝑥1 5
(5 − 4𝑝 − 3𝑞)
• 𝑥1 − 𝑥2 + 2𝑥3 + 𝑥4 = 2 , 𝑥 1
• 𝑥2 = 5
(−5 + 6𝑝 + 2𝑞) as p and q
3
• 5𝑥2 − 6𝑥3 − 2𝑥4 = −5 𝑝
𝑥4
• Let 𝑥3 = 𝑝 and 𝑥4 = 𝑞, an arbitrary 𝑞
constant for free variables varies.
• Substituting back, we get
1
• 𝑥2 = (−5 + 6𝑝 + 2𝑞) and
5
1
• 𝑥1 = (5 − 4𝑝 − 3𝑞)
5
• Hence the infinite values of solution are
given by,
 Example 5
• Investigate for what values of a and b the following linear • (i) For no solution:
equations
In this case 𝜌 𝐴 ≠ 𝜌 𝐴: 𝐵 .
• 𝑥 + 2𝑦 + 3𝑧 = 4, 𝑥 + 3𝑦 + 4𝑧 = 5, 𝑥 + 3𝑦 + 𝑎𝑧 = 𝑏, have
Then we have 𝜌 𝐴 = 2 𝑎𝑛𝑑 𝜌 𝐴: 𝐵 = 3.
(i) no solution, (ii) a unique solution,
i.e. 𝑎 = 4 and 𝑏 ≠ 5,
• (iii) Infinite number of solutions.
• (ii) For unique solution:
• Solution: The system of linear equations can be written in
the matrix form as 𝐴𝑋 = 𝐵 we should have 𝜌 𝐴 = 𝜌 𝐴: 𝐵 = 𝑛 = 3
1 2 3 𝑥 4
1 3 4 𝑦 = 5 In this case the system is consistent. Further,
1 3 𝑎 𝑧 𝑏 Since 𝜌 𝐴 = number of unknowns, therefore the system
1 2 3 4 possesses unique solution if 𝑎 ≠ 4 and for any value of b.
where the augmented matrix is 𝐴: 𝐵 = 1 3 4 5
• (iii) For infinite number of solutions:
1 3 𝑎 𝑏
Applying 𝑅2 − 𝑅1 𝑎𝑛𝑑 𝑅3 − 𝑅1 , we get we get𝜌 𝐴: 𝐵 = 𝜌 𝐴 = 2 < 3, the number of unknowns,
1 2 3 4 therefore system of equations is consistent and possesses
𝐴: 𝐵 ~ 0 1 1 1 an infinite number of solutions when 𝑎 = 4 and 𝑏 = 5,
0 1 𝑎−3 𝑏−4
1 2 3 4
Applying 𝑅3 − 𝑅2 , 𝐴: 𝐵 ~ 0 1 1 1
0 0 𝑎−4 𝑏−5
 Example 6
• For which values of  following set of equations is
consistent? Find and solve equations for those values
From above reduced form of 𝐴: 𝐵 it is clear that the
• 𝑥 + 2𝑦 + 𝑧 = 3, 𝑥 + 𝑦 + 𝑧 = 𝜆, 3𝑥 + 𝑦 + 3𝑧 = 𝜆2 rank of 𝐴 = 2. To have the rank of 𝐴: 𝐵 = 2,
1 2 1 3 • Consider 𝜆2 − 5𝜆 + 6 = 0 i. e. , 𝜆 − 2 𝜆 − 3 = 0
• Solution: consider 𝐴: 𝐵 = 1 1 1 𝜆 ⇒ 𝜆 = 2 𝑎𝑛𝑑 𝜆 = 3
3 1 3 𝜆2
• (i) Solution for 𝝀 = 𝟐
• Applying 𝑅2 − 𝑅1 𝑎𝑛𝑑 𝑅3 − 3𝑅1 , we get
1 2 1 3 • The reduced Echelon form of 𝐴: 𝐵 is
1 2 1 3
𝐴: 𝐵 ~ 0 −1 0 𝜆 − 3 0 −1 0 𝜆−3
0 −5 0 𝜆2 − 9 0 0 0 𝜆2 − 5𝜆 + 6
• Applying 𝑅3 − 5𝑅2 , we get 1 2 1 3
1 2 1 3 • Substituting 𝜆 = 2, we have 0 −1 0 −1
𝐴: 𝐵 ~ 0 −1 0 𝜆−3 0 0 0 0
0 0 0 𝜆2 − 5𝜆 + 6 • The reduced form of linear equations is
• For consistency of the equation, the rank of A and • 𝑥 + 2𝑦 + 𝑧 = 3 𝑎𝑛𝑑 𝑦 = 1
rank of 𝐴: 𝐵 must be the same.
 Example 6 (contd…)
• Let 𝑧 = 𝑘, an arbitrary constant, ∴ 𝑥 = 1 − 𝑘
𝑥 1−𝑘
• Hence 𝑦 = 1 has infinite values as k varies
𝑧 𝑘
• (ii) Solution for 𝝀 = 𝟑:
• The reduced Echelon form of 𝐴: 𝐵
1 2 1 3
• By substituting 𝜆 = 3, we have 0 −1 0 0
0 0 0 0
• The reduced form of linear equations is
• 𝑥 + 2𝑦 + 𝑧 = 3 𝑎𝑛𝑑 𝑦 = 0
• Let 𝑧 = 𝑐, an arbitrary constant. ∴ 𝑥 = 3 − 𝑐
𝑥 3−𝑐
• Hence 𝑦 = 0 has infinite solutions as c varies.
𝑧 𝑐
 Example 7
• For what values of 𝜆 the equations 3𝑥 − 2𝑦 + 𝜆𝑧 =
1, 2𝑥 + 𝑦 + 𝑧 = 2, 𝑥 + 2𝑦 − 𝜆𝑧 = −1, will have no • and they have no unique solutions if 𝜆 = −2
unique solution? Will the equations have any
solutions for this value of 𝜆. • Further, if 𝜆 = −2, we have from (1)
• Solution: (taking the equations in reverse order) 1 2 2 𝑥 −1
0 −3 −3 𝑦 = 4
1 2 −𝜆 𝑥 −1
0 −8 −8 𝑧 4
• 2 1 1 𝑦 = 2 8
3 −2 𝜆 𝑧 1 • Applying 𝑅3 − 𝑅2 , we get
3
• Applying 𝑅2 − 2𝑅1 𝑎𝑛𝑑 𝑅3 − 3𝑅1 , we get 1 2 2 𝑥 −1
1 2 −𝜆 𝑥 −1 0 −3 −3 𝑦 = 4
0 −3 1 + 2𝜆 𝑦 = 4 …………..(1) 0 0 0 𝑧 −20/3
0 −8 4𝜆 𝑧 4 • ∴ 0𝑥 + 0𝑦 + 0𝑧 = −20/3 which is absurd
• The equations have unique solutions if the coefficient
• Also the rank of A = 2 < 𝑡ℎ𝑒 𝑟𝑎𝑛𝑘 𝑜𝑓 [𝐴, 𝐵] = 3
matrix is non – singular.
• ∴ The equations are inconsistent, For 𝜆 = −2 there is
• ∴ −12𝜆 + 8 + 16𝜆 ≠ 0, 4𝜆 ≠ −8 ∴ 𝜆 ≠ −2
no solution.
• ∴ The equations have unique solutions if 𝜆 ≠ −2