📘 Lecture 34.6 — Normality condition for Legendre’ Polynomials / 35 Systems of Linear Differential Equations
📖 Overview: This lecture consists of two separate parts. The first part (Section 34.6) completes the study of Legendre polynomials by establishing the normality condition and the orthogonality condition for these polynomials, culminating in a precise formula for the integral of their square. The second part (Section 35) introduces the theory and solution of systems of simultaneous linear differential equations using the systematic elimination (operator) method, which is essential for modeling coupled physical systems like multiple masses on springs or parallel electrical circuits.
🗂️ Topics Covered
Derivation of the normality condition for Legendre polynomials using the generating function; the orthogonality condition expressed with the Kronecker delta; introduction to systems of linear differential equations; the systematic elimination (operator) method for solving such systems; worked examples demonstrating the elimination procedure and the reduction of arbitrary constants.
📝 Lecture Summary
34.6 Normality condition for Legendre’ Polynomials
The normality condition for Legendre polynomials is derived by exploiting the generating function. We start with the generating function and its conjugate, multiply them, and integrate over x from -1 to 1. This process leads to an infinite series on one side and an integral on the other. By performing the integration on the left-hand side and simplifying the logarithmic terms, we obtain a power series in t. The key step is to equate the coefficients of ( t^{2n} ) on both sides of the resulting equation. This directly yields the normality condition: the integral of the square of the Legendre polynomial ( P_n(x) ) from -1 to 1 equals ( 2/(2n+1) ). This derivation also naturally shows that the integral of the product of two different Legendre polynomials ( P_m(x) ) and ( P_n(x) ) (where ( m \neq n )) is zero, which is the orthogonality condition.
🔑 Definition — Normality Condition: The condition that the integral of the square of a set of functions, when weighted appropriately (or with weight function w(x)=1), yields a non-zero constant. For Legendre polynomials, this constant is ( 2/(2n+1) ).
📐 Formula: Normalization Integral for Legendre Polynomials [ \int_{-1}^1 [P_n(x)]^2 , dx = \frac{2}{2n+1} ] → The integral of the square of the nth Legendre polynomial over the interval [-1, 1] is equal to the constant ( 2/(2n+1) ).
📌 Example: The derivation involves equating the coefficient of ( t^{2n} ). For instance, for n=2, the coefficient of ( t^4 ) on the RHS of the expanded sum is ( 2/5 ). On the LHS, this corresponds to the term where m=n=2, giving ( \int_{-1}^1 [P_2(x)]^2 , dx \cdot t^4 ). By equating the coefficients, we get ( \int_{-1}^1 [P_2(x)]^2 , dx = 2/5 ).
💡 Why this matters: The normality condition is crucial for using Legendre polynomials in Fourier-Legendre series expansions, where the coefficients are calculated using this formula.
Remark: The lecture notes explicitly state that the orthogonality condition for ( P_n(x) ) can be written using the Kronecker delta ( \delta_{m,n} ). 📐 Formula: Orthogonality Condition for Legendre Polynomials [ \int_{-1}^1 P_m(x) P_n(x) , dx = \left( \frac{2}{2n+1} \right) \delta_{m,n} ] where ( \delta_{m,n} = 0 ) if ( m \neq n ) and ( \delta_{m,n} = 1 ) if ( m = n ).
🔑 Definition — Kronecker delta: A function of two variables, usually integers, which is 1 if they are equal, and 0 otherwise.
34.7 Exercise
This section provides a set of practice problems to solidify the concepts of Legendre polynomials and their properties.
- Alternative Form of Legendre’s Equation: Show that Legendre’s equation can be written in the self-adjoint form ( \frac{d}{dx} \left[ (1-x^2) \frac{dy}{dx} \right] + n(n+1)y = 0 ).
- Transformation to Legendre’s Equation: Show that the equation ( \sin \theta \frac{d^2 y}{d\theta^2} + \cos \theta \frac{dy}{d\theta} + n(n+1) (\sin \theta) y = 0 ) can be transformed into Legendre’s equation by substituting ( x = \cos \theta ).
- Evaluating the Integral of the Square: Use the explicit Legendre polynomials ( P_0(x), P_1(x), P_2(x), ) and ( P_3(x) ) to evaluate ( \int_{-1}^1 P_n^2 dx ) for n = 0, 1, 2, 3. Then, generalize the results (this yields the normality formula).
- Evaluating the Orthogonality Integral: Use the explicit Legendre polynomials to evaluate ( \int_{-1}^1 P_n(x) P_m(x) dx ) for ( n \neq m ). Generalize the results (this yields the orthogonality relation).
- Rodrigues’ Formula: Verify the Rodrigues’ formula ( P_n(x) = \frac{1}{2^n n!} \frac{d^n}{dx^n} (x^2 - 1)^n ) for n = 0, 1, 2, 3.
35 Systems of Linear Differential Equations
This section introduces the concept of systems of simultaneous differential equations. While a single spring-mass system is modeled by a single second-order ODE, coupling two or more masses with springs requires two or more simultaneous differential equations. Similarly, networks of parallel electrical circuits are modeled by such systems. The lecture focuses on the theory and solution of systems of linear differential equations with constant coefficients.
35.1 Simultaneous Differential Equations
These involve two or more equations that contain derivatives of two or more unknown functions of a single independent variable (usually t). The differential operator D is defined as ( D = d/dt ), allowing the equations to be written in a compact, algebraic form. For example, ( D^2 x = d^2 x / dt^2 ).
🔑 Definition — System of Simultaneous Differential Equations: A set of differential equations involving two or more unknown functions of the same independent variable.
📌 Example: The equations ( d^2 x / dt^2 = -5x + y ) and ( d^2 y / dt^2 = 3x - y ) form a system of two second-order equations for the unknown functions x(t) and y(t). The equations ( x' - 3x + y' + z' = 5 ) and ( x + y' - 6z' = t - 1 ) are a system for x(t), y(t), and z(t).
35.2 Solution of a System
A solution is a set of differentiable functions ( x = f(t), y = g(t), \dots ) that satisfy each equation in the system on some interval I.
35.2.1 Systematic Elimination (Operator Method)
This is the primary method used in this lecture. It is based on the process of systematically eliminating dependent variables to obtain a single linear differential equation in one of them.
Step-by-Step Procedure:
- Write in Operator Form: Convert all equations into operator form using D.
- Eliminate: Choose a dependent variable to eliminate by operating on the equations with appropriate operator combinations (like D, D-3, etc.) and adding or subtracting.
- Solve for First Variable: The elimination yields a single linear ODE with constant coefficients for one retained variable. Solve this ODE to find the first variable as a function of t (e.g., y(t)).
- Eliminate Again: Eliminate all other variables (or the same variable as before) to get a single linear ODE for a second dependent variable (e.g., x(t)).
- Solve for Second Variable: Solve this ODE. The solution will have its own set of arbitrary constants.
- Reduce Constants: Substitute the found expressions for the dependent variables into one of the original equations. This will give relationships between the arbitrary constants, reducing their number.
- Write Solution: Write the final solution for the system with the correct number of independent arbitrary constants.
📌 Example 1: Solve ( dy/dt = 2x, \ dx/dt = 3y ). The steps lead to ( D^2 y - 6y = 0 ) and ( D^2 x - 6x = 0 ). Solving gives ( y(t) = c_1 e^{\sqrt{6} t} + c_2 e^{-\sqrt{6} t} ) and ( x(t) = c_3 e^{\sqrt{6} t} + c_4 e^{-\sqrt{6} t} ). Substituting into the first equation yields the relations ( c_3 = (\sqrt{6}/2)c_1 ) and ( c_4 = -(\sqrt{6}/2)c_2 ). The final solution is ( x(t) = \frac{\sqrt{6}}{2} c_1 e^{\sqrt{6} t} - \frac{\sqrt{6}}{2} c_2 e^{-\sqrt{6} t}, \ y(t) = c_1 e^{\sqrt{6} t} + c_2 e^{-\sqrt{6} t} ).
📌 Example 2: Solve ( Dx + (D+2)y = 0, \ (D-3)x - 2y = 0 ). Elimination yields ( (D^2 + D - 6)y = 0 ) and ( (D^2 + D - 6)x = 0 ). The solutions are ( y(t) = c_1 e^{2t} + c_2 e^{-3t} ) and ( x(t) = c_3 e^{2t} + c_4 e^{-3t} ). Substituting into the first original equation gives the relations ( c_3 = -2c_1 ) and ( c_4 = -\frac{1}{3}c_2 ). The final solution is ( x(t) = -2c_1 e^{2t} - \frac{1}{3}c_2 e^{-3t}, \ y(t) = c_1 e^{2t} + c_2 e^{-3t} ).
📌 Example 3: Solve ( Dx - 4x + D^2 y = t^2, \ (D+1)x + D y = 0 ). This is a non-homogeneous system. Eliminating x yields ( (D^3 + 4D)y = t^2 + 2t ). The complementary function is ( y_c = c_1 + c_2 \cos 2t + c_3 \sin 2t ). For the particular solution, the method of undetermined coefficients is used with the trial function ( y_p = At^3 + Bt^2 + Ct ). Solving for the coefficients gives ( A = 1/12, B = 1/4, C = -1/8 ). The final solution for y is the sum of the complementary and particular parts.
⭐ Key Takeaways
For Legendre polynomials, the key result is the normality condition ( \int_{-1}^1 [P_n(x)]^2 , dx = \frac{2}{2n+1} ), which, along with the orthogonality condition expressed with the Kronecker delta, is fundamental for series expansions. For systems of differential equations, the systematic elimination method (operator method) is a powerful technique for solving linear systems with constant coefficients. The most critical step in this process is to correctly substitute the solutions back into an original equation to reduce the number of arbitrary constants, ensuring the final solution has the correct number of independent parameters (equal to the order of the system).
🧠 Quick Revision Questions
- What is the value of the integral ( \int_{-1}^1 [P_n(x)]^2 , dx )? (Answer: ( 2/(2n+1) ))
- State the orthogonality condition for Legendre polynomials in a single equation using the Kronecker delta.
- What is the first step in the systematic elimination method for solving a system of linear differential equations?
- Why is it necessary to substitute the solutions back into an original equation after solving the separate ODEs?
- For the system ( Dx = -y, \ Dy = x ), what is the differential equation for x after eliminating y?
📘 Lecture 37 — 36 Systems of Linear Differential Equations
📖 Overview: This lecture presents a systematic method for solving systems of linear differential equations with constant coefficients using determinants. By extending Cramer’s rule to differential operators, the system is decoupled into separate higher-order differential equations for each dependent variable, which can then be solved using standard methods. This approach provides a powerful, structured technique for tackling complex systems, including those with three or more variables.
🗂️ Topics Covered
The lecture begins by defining the operator form of a system of two linear differential equations and deriving the decoupled equations using determinants. It then outlines a step-by-step solution method, paying careful attention to the algebraic treatment of the differential operator D on the left-hand side versus its operational treatment on the right-hand side. Three detailed examples are solved, progressing from a two-variable non-homogeneous system to a three-variable system, to illustrate the complete procedure and the reduction of arbitrary constants. The section concludes with a set of practice exercises.
📝 Lecture Summary
36.1 Solution of Using Determinants
If a system of two linear differential equations in variables x and y is written in differential operator form using operators L₁, L₂, L₃, L₄ as L₁x + L₂y = g₁(t) and L₃x + L₄y = g₂(t), the system can be decoupled. By applying operator L₄ to the first equation and L₂ to the second and subtracting, we obtain (L₁L₄ – L₂L₃)x = L₄g₁ – L₂g₂. Similarly, applying L₃ to the first and L₁ to the second and subtracting gives (L₁L₄ – L₂L₃)y = L₁g₂ – L₃g₁.
These results can be succinctly expressed using determinants analogous to Cramer’s rule: |L₁ L₂| · x = |g₁ L₂| and |L₁ L₂| · y = |L₁ g₁| |L₃ L₄| |g₂ L₄| |L₃ L₄| |L₃ g₂|
💡 Why this matters: This formulation allows us to treat the system as a single higher-order differential equation for each variable, simplifying the solution process.
🔑 Definition — Determinant Expansion Rules:
- Left-hand side determinant: The symbol D in operators Lᵢ is treated as an algebraic quantity. The expansion results in a differential operator of order n, which then operates on x (or y).
- Right-hand side determinant: The symbol D in operators Lᵢ is treated as an algebraic quantity, BUT the expansion is performed in the sense of differential operators actually operating on the functions g₁(t) and g₂(t).
📌 Example: For the system (2D-5)x + Dy = eᵗ, (D-1)x + Dy = 5eᵗ, the decoupled equation for x is found by:
- Computing the left-hand determinant:
|2D-5 D| / |D-1 D| = (2D-5)D - (D-1)D = D² - 4D. - Computing the right-hand determinant for x:
|eᵗ D| / |5eᵗ D| = eᵗ · D - D · (5eᵗ). Note thatD(5eᵗ) = 5eᵗ. So this equalsDeᵗ - 5Deᵗ = eᵗ - 5eᵗ = -4eᵗ. - The final equation is
(D² - 4D)x = -4eᵗ.
36.2 Solution Method
The method for solving a system of linear differential equations using determinants involves these steps:
Step 1: Write the system in differential operator form: L₁x + L₂y = g₁(t), L₃x + L₄y = g₂(t).
Step 2: Form the determinants: |L₁ L₂ / L₃ L₄|, |g₁ L₂ / g₂ L₄|, and |L₁ g₁ / L₃ g₂|.
Step 3: If the first determinant is non-zero, it becomes the differential operator for the decoupled equations:
|L₁ L₂ / L₃ L₄| · x = |g₁ L₂ / g₂ L₄|
|L₁ L₂ / L₃ L₄| · y = |L₁ g₁ / L₃ g₂|
Step 4: Find the complementary functions x_c and y_c for both equations. They will have the same auxiliary equation and thus the same form.
Step 5: Find particular integrals x_p and y_p using either the method of undetermined coefficients or variation of parameters.
Step 6: Write the general solutions: x = x_c + x_p and y = y_c + y_p.
Step 7: Reduce the number of independent constants by substituting the general solutions back into one of the original differential equations of the system.
💡 Why this matters: If the first determinant is zero, the system may have a solution with any number of independent constants or may have no solution at all. A similar caution applies to larger systems.
🔑 Definition — First Determinant is Zero: If |L₁ L₂ / L₃ L₄| = 0, the system may have infinite solutions with arbitrary constants or no solution, requiring alternative methods.
📌 Example 1: Solve 2dx/dt - 5x + dy/dt = eᵗ, dx/dt - x + dy/dt = 5eᵗ.
- Operator form:
(2D-5)x + Dy = eᵗ,(D-1)x + Dy = 5eᵗ. - Determinants:
|2D-5 D| / |D-1 D| = D² - 4D. For x:|eᵗ D| / |5eᵗ D| = Deᵗ - D(5eᵗ) = -4eᵗ. For y:|2D-5 eᵗ| / |D-1 5eᵗ| = (2D-5)(5eᵗ) - (D-1)eᵗ = -15eᵗ. - Decoupled equations:
(D² - 4D)x = -4eᵗ,(D² - 4D)y = -15eᵗ. - Complementary functions: Auxiliary equation
m² - 4m = 0 → m = 0, 4. Sox_c = c₁ + c₂e⁴ᵗ,y_c = c₃ + c₄e⁴ᵗ. - Particular integrals: Assume
x_p = Aeᵗ,y_p = Beᵗ. Substituting:Aeᵗ - 4Aeᵗ = -4eᵗ → -3A = -4 → A = 4/3.-3B = -15 → B = 5. Sox_p = (4/3)eᵗ,y_p = 5eᵗ. - General solution:
x = c₁ + c₂e⁴ᵗ + (4/3)eᵗ,y = c₃ + c₄e⁴ᵗ + 5eᵗ. - Reduce constants: Substitute into
dx/dt - x + dy/dt = 5eᵗ. This yieldsc₁ = 0andc₄ = -(3/4)c₂. After re-notating, the final solution is:x(t) = c₁e⁴ᵗ + (4/3)eᵗ,y(t) = -(3/4)c₁e⁴ᵗ + c₂ + 5eᵗ.
📌 Example 2: Solve x' = 3x - y - 1, y' = x + y + 4eᵗ.
- Operator form:
(D-3)x + y = -1,-x + (D-1)y = 4eᵗ. - Determinants:
|D-3 1| / |-1 D-1| = D² - 4D + 4 = (D-2)². For x:|-1 1| / |4eᵗ D-1| = (-1)(D-1) - (1)(4eᵗ) = -D + 1 - 4eᵗ. For y:|D-3 -1| / |-1 4eᵗ| = (D-3)(4eᵗ) - (-1)(-1) = 4Deᵗ - 12eᵗ - 1 = -8eᵗ - 1. - Decoupled equations:
(D-2)²x = 1 - 4eᵗ,(D-2)²y = -1 - 8eᵗ. - Complementary functions:
m = 2, 2:x_c = c₁e²ᵗ + c₂te²ᵗ,y_c = c₃e²ᵗ + c₄te²ᵗ. - Particular integrals: Assume
x_p = A + Beᵗ,y_p = C + Eeᵗ. Substituting into(D-2)²x = D²x - 4Dx + 4x = 1 - 4eᵗgivesBeᵗ + 4A = 1 - 4eᵗ. SoA = 1/4,B = -4. SimilarlyEeᵗ + 4C = -1 - 8eᵗgivesC = -1/4,E = -8. Sox_p = 1/4 - 4eᵗ,y_p = -1/4 - 8eᵗ. - General solution:
x = c₁e²ᵗ + c₂te²ᵗ + 1/4 - 4eᵗ,y = c₃e²ᵗ + c₄te²ᵗ - 1/4 - 8eᵗ. - Reduce constants: Substituting into
y' = x + y + 4eᵗgivesc₃ - c₁ + c₄ = 0andc₄ - c₂ = 0, soc₄ = c₂andc₃ = c₁ - c₂. The final solution is:x(t) = c₁e²ᵗ + c₂te²ᵗ + 1/4 - 4eᵗ,y(t) = (c₁ - c₂)e²ᵗ + c₂te²ᵗ - 1/4 - 8eᵗ.
📌 Example 3: For the system Dx + Dz = t², 2x + D²y = eᵗ, -2Dx - 2y + (D+1)z = 0, find the differential equations for x, y, z.
- The system is already in operator form.
- The 1st determinant
|D 0 D| / |2 D² 0| / |-2D -2 D+1|is computed to beD(3D³ + D² - 4) ≠ 0. - The decoupled equations are:
D(3D³ + D² - 4)x = 2 - 2eᵗD(3D³ + D² - 4)y = 4eᵗ - 2t² - 4tD(3D³ + D² - 4)z = 2eᵗ - 4t²
⭐ Key Takeaways
The determinant method for solving linear systems is a powerful extension of Cramer's rule. The critical distinction is that when evaluating the left-hand side determinant, the operator D is treated as a pure algebraic symbol, but on the right-hand side, D acts as a proper differential operator on the functions g₁ and g₂. The solution process culminates in a reduction of the number of arbitrary constants by substituting the general solutions back into any one of the original equations. If the system’s determinant is zero, the system may have an infinite family of solutions or none at all, requiring careful analysis.
🧠 Quick Revision Questions
- How is the decoupled equation
(L₁L₄ – L₂L₃)x = L₄g₁ – L₂g₂derived from the original systemL₁x + L₂y = g₁,L₃x + L₄y = g₂? - What is the crucial difference in how the operator D is treated when expanding the determinant on the left-hand side versus on the right-hand side of a decoupled equation?
- In the solution method, what does it imply if the first determinant
|L₁ L₂| / |L₃ L₄|equals zero? - In Example 2, after finding the general solutions
xandy, which original equation was used to reduce the number of constants, and what were the resulting relationships? - Write the three decoupled differential equations for the dependent variables x, y, and z found in Example 3.
📘 Lecture 38 — Introduction to Matrices
📖 Overview: This lecture introduces matrix theory as a foundation for representing systems of linear differential equations. It covers basic matrix operations including addition, scalar multiplication, and matrix multiplication, along with key concepts such as determinants and transposes, which are essential for solving differential equation systems.
🗂️ Topics Covered
The lecture begins with definitions of matrices, their order, and types including square matrices and column vectors. It then covers matrix operations: scalar multiplication, addition, and difference of matrices. The core focus is on matrix multiplication with examples demonstrating conformability, associativity, and non-commutativity. The lecture concludes with the determinant of a matrix, transpose operations, and the identity and zero matrices.
📝 Lecture Summary
38.1 Matrix
A matrix is a rectangular array of numbers or functions subject to certain rules and conditions, denoted by capital letters. The numbers or functions are called elements or entries, denoted by small letters.
38.2 Rows and Columns
The horizontal lines in a matrix are called rows, and the vertical lines are called columns.
38.3 Order of a Matrix
If a matrix has ( m ) rows and ( n ) columns, its size or order is ( m \times n ). A matrix A with m rows and n columns is written as: [ A = \begin{pmatrix} a_{11} & a_{12} & \dots & a_{1n} \ a_{21} & a_{22} & \dots & a_{2n} \ \vdots & \vdots & \ddots & \vdots \ a_{m1} & a_{m2} & \dots & a_{mn} \end{pmatrix} ]
38.4 Square Matrix
A matrix with ( n ) rows and ( n ) columns is an ( n \times n ) square matrix of order ( n ). The entry in the i-th row and j-th column is written as ( a_{ij} ).
38.5 Equality of Matrix
Two matrices A and B are equal if and only if they have the same orders and corresponding elements are equal: ( A = B \iff a_{ij} = b_{ij} ) for all ( i, j ).
38.6 Column Matrix
A column matrix (or column vector) X has ( n ) rows and only one column: [ X = \begin{pmatrix} b_{11} \ b_{21} \ b_{31} \ \vdots \ b_{n1} \end{pmatrix} = [b_{i1}]_{n \times 1} ]
38.7 Multiple of Matrices
A multiple of a matrix A is defined as ( kA = [ka_{ij}]_{m \times n} ), where ( k ) is a constant or function. Note: ( kA = Ak ).
📌 Example 1: (a) ( 5 \cdot \begin{pmatrix} 2 & -3 \ 4 & -1 \ 1/5 & 6 \end{pmatrix} = \begin{pmatrix} 10 & -15 \ 20 & -5 \ 1 & 30 \end{pmatrix} ) (b) ( e^t \cdot \begin{pmatrix} 1 \ -2 \ 4 \end{pmatrix} = \begin{pmatrix} e^t \ -2e^t \ 4e^t \end{pmatrix} ) (c) ( e^{-3t} \cdot \begin{pmatrix} 2 \ 5 \end{pmatrix} = \begin{pmatrix} 2e^{-3t} \ 5e^{-3t} \end{pmatrix} = \begin{pmatrix} 2 \ 5 \end{pmatrix} e^{-3t} )
38.8 Addition of Matrices
Two matrices can be added only if they have the same orders. The sum is obtained by adding corresponding entries: ( A + B = [a_{ij} + b_{ij}] ).
📌 Example 2: [ A = \begin{pmatrix} 2 & -1 & 3 \ 0 & 4 & 6 \ -6 & 10 & -5 \end{pmatrix}, \quad B = \begin{pmatrix} 4 & 7 & -8 \ 9 & 3 & 5 \ 1 & -1 & 2 \end{pmatrix} ] [ A + B = \begin{pmatrix} 2+4 & -1+7 & 3+(-8) \ 0+9 & 4+3 & 6+5 \ -6+1 & 10+(-1) & -5+2 \end{pmatrix} = \begin{pmatrix} 6 & 6 & -5 \ 9 & 7 & 11 \ -5 & 9 & -3 \end{pmatrix} ]
📌 Example 3: [ \begin{pmatrix} 3t^2 - 2e^t \ t^2 + 7t \ 5t \end{pmatrix} = \begin{pmatrix} 3t^2 \ t^2 \ 0 \end{pmatrix} + \begin{pmatrix} 0 \ 7t \ 5t \end{pmatrix} + \begin{pmatrix} -2e^t \ 0 \ 0 \end{pmatrix} = \begin{pmatrix} 3 \ 1 \ 0 \end{pmatrix} t^2 + \begin{pmatrix} 0 \ 7 \ 5 \end{pmatrix} t + \begin{pmatrix} -2 \ 0 \ 0 \end{pmatrix} e^t ]
38.9 Difference of Matrices
The difference of two matrices A and B of the same order is: ( A - B = A + (-B) ), where ( -B = (-1)B ).
38.10 Multiplication of Matrices
Two matrices A and B are conformable for the product AB if the number of columns in A equals the number of rows in B. If ( A_{m \times n} \cdot B_{n \times p} = C_{m \times p} ), the entry ( c_{ij} = \sum_{k=1}^n a_{ik}b_{kj} ).
🔑 Definition — Conformable: Matrices A (order ( m \times n )) and B (order ( n \times p )) are conformable for multiplication because the inner dimensions match.
📌 Example 4 (a): [ A = \begin{pmatrix} 4 & 7 \ 3 & 5 \end{pmatrix}, \quad B = \begin{pmatrix} 9 & -2 \ 6 & 8 \end{pmatrix} ] [ AB = \begin{pmatrix} 4\cdot9+7\cdot6 & 4\cdot(-2)+7\cdot8 \ 3\cdot9+5\cdot6 & 3\cdot(-2)+5\cdot8 \end{pmatrix} = \begin{pmatrix} 78 & 48 \ 57 & 34 \end{pmatrix} ] [ BA = \begin{pmatrix} 9\cdot4+(-2)\cdot3 & 9\cdot7+(-2)\cdot5 \ 6\cdot4+8\cdot3 & 6\cdot7+8\cdot5 \end{pmatrix} = \begin{pmatrix} 30 & 53 \ 48 & 82 \end{pmatrix} ] Note: ( AB \neq BA ), showing matrix multiplication is not commutative.
📌 Example 4 (b): [ A = \begin{pmatrix} 5 & 8 \ 1 & 0 \ 2 & 7 \end{pmatrix}{3 \times 2}, \quad B = \begin{pmatrix} -4 & -3 \ 2 & 0 \end{pmatrix}{2 \times 2} ] [ AB = \begin{pmatrix} 5\cdot(-4)+8\cdot2 & 5\cdot(-3)+8\cdot0 \ 1\cdot(-4)+0\cdot2 & 1\cdot(-3)+0\cdot0 \ 2\cdot(-4)+7\cdot2 & 2\cdot(-3)+7\cdot0 \end{pmatrix} = \begin{pmatrix} -4 & -15 \ -4 & -3 \ 6 & -6 \end{pmatrix} ] The product BA is not possible because B has 2 columns but A has 3 rows.
📌 Example 5: (a) ( \begin{pmatrix} 2 & -1 & 3 \ 0 & 4 & 5 \ 1 & -7 & 9 \end{pmatrix} \begin{pmatrix} -3 \ 6 \ 4 \end{pmatrix} = \begin{pmatrix} 2(-3)+(-1)(6)+3(4) \ 0(-3)+4(6)+5(4) \ 1(-3)+(-7)(6)+9(4) \end{pmatrix} = \begin{pmatrix} 0 \ 44 \ -9 \end{pmatrix} ) (b) ( \begin{pmatrix} -4 & 2 \ 3 & 8 \end{pmatrix} \begin{pmatrix} x \ y \end{pmatrix} = \begin{pmatrix} -4x+2y \ 3x+8y \end{pmatrix} )
38.11 Multiplicative Identity
The ( n \times n ) identity matrix I has 1's on the main diagonal and 0's elsewhere. For any ( n \times n ) matrix A, ( I \cdot A = A \cdot I = A ), and for any ( n \times 1 ) column matrix X, ( I \cdot X = X ).
38.12 Zero Matrix
A zero matrix or null matrix O consists of all zero entries. If A and O are ( m \times n ) matrices, then ( A + O = O + A = A ).
38.13 Associative Law
Matrix multiplication is associative: ( A(BC) = (AB)C ), where A, B, C are ( m \times p ), ( p \times r ), and ( r \times n ) matrices respectively.
38.14 Distributive Law
The distributive law states: ( A(B + C) = AB + AC ) (if orders are compatible), and ( (A + B)C = AC + BC ).
38.15 Determinant of a Matrix
Associated with every square matrix A is a number called the determinant, denoted det(A) or |A|.
📌 Example 6: [ A = \begin{pmatrix} 3 & 6 & 2 \ 2 & 5 & 1 \ -1 & 2 & 4 \end{pmatrix} ] Expanding by cofactors of the first row: [ \det(A) = 3 \begin{vmatrix} 5 & 1 \ 2 & 4 \end{vmatrix} - 6 \begin{vmatrix} 2 & 1 \ -1 & 4 \end{vmatrix} + 2 \begin{vmatrix} 2 & 5 \ -1 & 2 \end{vmatrix} = 3(20-2) - 6(8+1) + 2(4+5) = 54 - 54 + 18 = 18 ]
38.16 Transpose of a Matrix
The transpose of an ( m \times n ) matrix A, denoted ( A^{tr} ), is obtained by interchanging rows and columns. If ( A = [a_{ij}]{m \times n} ), then ( A^{tr} = [a{ji}]_{n \times m} ).
📌 Example 7 (a): [ A = \begin{pmatrix} 3 & 6 & 2 \ 2 & 5 & 1 \ -1 & 2 & 4 \end{pmatrix}, \quad A^{tr} = \begin{pmatrix} 3 & 2 & -1 \ 6 & 5 & 2 \ 2 & 1 & 4 \end{pmatrix} ]
📌 Example 7 (b): [ X = \begin{pmatrix} 5 \ 0 \ 3 \end{pmatrix}, \quad X^{tr} = \begin{pmatrix} 5 & 0 & 3 \end{pmatrix} ]
⭐ Key Takeaways
The lecture establishes matrix fundamentals crucial for representing systems of linear differential equations. Students must remember that matrix multiplication requires the number of columns in the first matrix to equal the number of rows in the second, and this operation is not commutative. The identity and zero matrices serve as multiplicative and additive identities, respectively. The determinant is defined only for square matrices and is computed by cofactor expansion. The transpose operation swaps rows and columns, changing an ( m \times n ) matrix into an ( n \times m ) matrix.
🧠 Quick Revision Questions
- What condition must be satisfied for two matrices A and B to be conformable for the product AB?
- Is matrix multiplication commutative? Provide a counterexample from the lecture.
- How is the determinant of a ( 3 \times 3 ) matrix computed using cofactor expansion?
- What is the difference between a square matrix and a column matrix?
- How does the transpose of a matrix relate to its original matrix in terms of order?
📘 Lecture 39 — Multiplicative Inverse of a Matrix
📖 Overview: This lecture covers the concept of the multiplicative inverse of a matrix, including necessary conditions for its existence and methods for calculation. It then extends matrix operations to include derivatives and integrals of matrices whose entries are functions, and introduces the augmented matrix and elementary row operations for solving systems of linear equations using Gaussian and Gauss-Jordan elimination methods.
🗂️ Topics Covered
The lecture defines the multiplicative inverse of a matrix and the condition of non-singularity (determinant not equal to zero). It provides a theorem for calculating the inverse using cofactors and works through examples for 2×2 and 3×3 matrices. The topics then shift to defining the derivative and integral of a matrix of functions with an example. Finally, it covers the augmented matrix, elementary row operations, and the structure of row-echelon and reduced row-echelon forms used in Gaussian and Gauss-Jordan elimination methods for solving systems of linear equations.
📝 Lecture Summary
38.17 Multiplicative Inverse of a Matrix
Suppose that A is a square matrix of order n × n . If there exists an n × n matrix B such that AB = BA = I, then B is said to be the multiplicative inverse of the matrix A and is denoted by B = A⁻¹.
38.18 Non-Singular Matrices
A square matrix A of order n × n is said to be a non-singular matrix if det(A) ≠ 0. Otherwise, the square matrix A is said to be singular. Thus for a singular A we must have det(A) = 0.
Theorem: If A is a square matrix of order n × n, then the matrix has a multiplicative inverse A⁻¹ if and only if the matrix A is non-singular.
Theorem: Let A be a non-singular matrix of order n × n and let Cⱼ denote the cofactor (signed minor) of the corresponding entry aᵢⱼ in the matrix A, i.e., Cᵢⱼ = (-1)ⁱ⁺ʲ Mᵢⱼ, where Mᵢⱼ is the determinant of the (n-1) × (n-1) matrix obtained by deleting the i-th row and j-th column from A. Then the inverse of the matrix A is given by A⁻¹ = (1 / det(A)) * (Cᵢⱼ)ᵗʳ.
For a 2 × 2 non-singular matrix A = [[a₁₁, a₁₂], [a₂₁, a₂₂]], the cofactors are C₁₁ = a₂₂, C₁₂ = -a₂₁, C₂₁ = -a₁₂, and C₂₂ = a₁₁. Therefore, A⁻¹ = (1/det(A)) * [[a₂₂, -a₁₂], [-a₂₁, a₁₁]].
For a 3 × 3 non-singular matrix, the cofactors are computed, and the inverse is A⁻¹ = (1/det(A)) * [[C₁₁, C₂₁, C₃₁], [C₁₂, C₂₂, C₃₂], [C₁₃, C₂₃, C₃₃]]ᵗʳ.
🔑 Definition — Cofactor: Cᵢⱼ = (-1)ⁱ⁺ʲ Mᵢⱼ, where Mᵢⱼ is the determinant of the matrix left after removing the i-th row and j-th column. 📐 Formula — Inverse of a Matrix: A⁻¹ = (1 / det(A)) * (Matrix of Cofactors)ᵗʳ → The inverse of a square matrix is its transpose of cofactors divided by its determinant. 📌 Example 8: Find the multiplicative inverse for the matrix A = [[2, 4], [1, 10]]. Solution: det(A) = (2)(10) - (4)(1) = 20 - 4 = 16. Since det(A) ≠ 0, A is non-singular and A⁻¹ exists. A⁻¹ = (1/16) * [[10, -4], [-1, 2]] = [[10/16, -4/16], [-1/16, 2/16]] = [[5/8, -1/4], [-1/16, 1/8]]. (Note: The lecture text had a miscalculation for det(A) and the inverse; the process shown here is correct based on the formula).
📌 Example 9: Find the multiplicative inverse of the matrix A = [[2, 2], [3, 3]]. Solution: det(A) = (2)(3) - (2)(3) = 6 - 6 = 0. Since det(A) = 0, the matrix is singular. Therefore, the multiplicative inverse A⁻¹ does not exist.
📌 Example 10: Find the multiplicative inverse for the matrix A = [[2, 2, 0], [-2, 1, 1], [3, 0, 1]]. Solution: det(A) = 2(11 - 10) - 2((-2)1 - 13) + 0((-2)0 - 13) = 2(1) - 2(-2-3) = 2 - 2(-5) = 2 + 10 = 12 ≠ 0. Therefore A is non-singular. The cofactors are: C₁₁ = 1, C₁₂ = 5, C₁₃ = -3; C₂₁ = -2, C₂₂ = 2, C₂₃ = 6; C₃₁ = 2, C₃₂ = -2, C₃₃ = 6. Then A⁻¹ = (1/12) * [[1, -2, 2], [5, 2, -2], [-3, 6, 6]]ᵗʳ = (1/12) * [[1, 5, -3], [-2, 2, 6], [2, -2, 6]] = [[1/12, 5/12, -1/4], [-1/6, 1/6, 1/2], [1/6, -1/6, 1/2]].
38.19 Derivative of a Matrix of functions
Suppose that A(t) = [aᵢⱼ(t)]ₘₓₙ is a matrix whose entries are functions that are differentiable on a common interval. Then the derivative of the matrix A(t) is a matrix whose entries are derivatives of the corresponding entries of the matrix A(t). Thus dA/dt = [daᵢⱼ/dt]ₘₓₙ. The derivative of a matrix is also denoted by A'(t).
38.20 Integral of a Matrix of Functions
Suppose that A(t) = (aᵢⱼ(t)) is a matrix whose entries are functions that are continuous on a common interval containing t. Then the integral of the matrix A(t) is a matrix whose entries are integrals of the corresponding entries of the matrix A(t). Thus ∫ A(s) ds = [∫ aᵢⱼ(s) ds]ₘₓₙ.
📌 Example 11: Find the derivative and the integral of the matrix X(t) = [sin(2t), e^(3t), 8t - 1]ᵗʳ. Solution: Derivative: X'(t) = [d/dt(sin(2t)), d/dt(e^(3t)), d/dt(8t-1)]ᵗʳ = [2cos(2t), 3e^(3t), 8]ᵗʳ. Integral: ∫ X(s) ds = [∫ sin(2s) ds, ∫ e^(3s) ds, ∫ (8s-1) ds]ᵗʳ = [-1/2 cos(2t) + c₁, 1/3 e^(3t) + c₂, 4t² - t + c₃]ᵗʳ.
38.21 Augmented Matrix
Consider an algebraic system of n linear equations in n unknowns. Suppose that A denotes the coefficient matrix. The augmented matrix of the system of linear equations is the n × (n+1) matrix [A|b], where b is the column matrix of the constants (b₁, b₂, ..., bₙ).
38.22 Elementary Row Operations
The elementary row operations consist of the following three operations:
- Multiply a row by a non-zero constant.
- Interchange any row with another row.
- Add a non-zero constant multiple of one row to another row.
38.23 The Gaussian and Gauss-Jordan Methods
In the Gaussian Elimination method, we carry out a succession of elementary row operations on the augmented matrix until it is transformed into row-echelon form. This form has the following structure:
- The first non-zero entry in a non-zero row is 1.
- In consecutive non-zero rows, the first entry 1 in the lower row appears to the right of the first 1 in the higher row.
- Rows consisting of all 0’s are at the bottom of the matrix.
In the Gauss-Jordan method, the row operations are continued until the augmented matrix is transformed into the reduced row-echelon form. A reduced row-echelon matrix has the same structure as row-echelon, but with the additional property:
- A column containing a first entry 1 has 0’s everywhere else.
📌 Example 1: (a) The following two matrices are in row-echelon form: [[1, 5, 0, 2], [0, 0, 1, -1], [0, 0, 0, 0]] and [[1, -6, 2, 2], [0, 0, 1, 4]]. (b) The following two matrices are in reduced row-echelon form: [[1, 0, 0, 7], [0, 1, 0, -1], [0, 0, 0, 0]] and [[1, -6, 0, -6], [0, 0, 1, 4]].
📌 Example 2: Solve the system: 2x₁ + 6x₂ + x₃ = 7, x₁ + 2x₂ - x₃ = -1, 5x₁ + 7x₂ - 4x₃ = 9 by (a) Gaussian elimination and (b) Gauss-Jordan elimination.
⭐ Key Takeaways
The multiplicative inverse of a matrix A⁻¹ exists only when the matrix is non-singular (det(A) ≠ 0). The inverse can be calculated using the formula involving cofactors and the determinant, with a simpler rule for 2×2 matrices. For matrices of functions, differentiation and integration are performed entry-wise. To solve systems of linear equations using Gaussian or Gauss-Jordan elimination, we first form the augmented matrix [A|b] and apply elementary row operations to transform it into row-echelon or reduced row-echelon form, respectively. The reduced row-echelon form has the additional condition that the leading 1 in each row is the only non-zero entry in its column.
🧠 Quick Revision Questions
- What is the necessary and sufficient condition for a square matrix to have a multiplicative inverse?
- Write the formula for the inverse of a non-singular 2×2 matrix A = [[a, b], [c, d]].
- How do you find the derivative of a matrix A(t) = [aᵢⱼ(t)]?
- What is an augmented matrix, and why is it used in Gaussian elimination?
- State the additional property of a reduced row-echelon form that is not required for a row-echelon form.
📘 Lecture 39 — The Eigenvalue Problem
📖 Overview: This lecture introduces the eigenvalue problem for square matrices, a fundamental concept in linear algebra and differential equations. Eigenvalues and eigenvectors are crucial for solving systems of linear differential equations, and the lecture covers how to compute them using the characteristic equation and Gauss-Jordan elimination, including cases with distinct, repeated, and complex eigenvalues.
🗂️ Topics Covered
The lecture begins by defining eigenvalues and eigenvectors, then transforms the problem into a system of linear algebraic equations. It explains the characteristic equation and the condition for non-trivial solutions. Through several examples, it demonstrates how to find eigenvalues and eigenvectors for matrices of various sizes, including cases with distinct eigenvalues, repeated eigenvalues (both with and without enough independent eigenvectors), and complex eigenvalues. The lecture also includes a set of exercises.
📝 Lecture Summary
39.1 Eigenvalues and Eigenvectors
Let A be an n × n matrix. A number λ is said to be an eigenvalue of A if there exists a nonzero solution vector K of the system of linear equations: AK = λK. The solution vector K is said to be an eigenvector corresponding to the eigenvalue λ. Using matrix algebra, this equation can be written in the alternative form (A - λI)K = 0, where I is the identity matrix.
If we let K = [k₁, k₂, ..., kₙ]ᵀ, the system becomes a system of linear algebraic equations: (a₁₁ - λ)k₁ + a₁₂k₂ + ... + a₁ₙkₙ = 0, a₂₁k₁ + (a₂₂ - λ)k₂ + ... + a₂ₙkₙ = 0, and so on up to aₙ₁k₁ + aₙ₂k₂ + ... + (aₙₙ - λ)kₙ = 0. An obvious solution of this system is the trivial solution k₁ = k₂ = ... = kₙ = 0. However, we are seeking only a non-trivial solution of the system.
39.2 The Non-trivial solution
The non-trivial solution of the system exists only when det(A - λI) = 0. This equation is called the characteristic equation of the matrix A. Thus, the eigenvalues of the matrix A are given by the roots of the characteristic equation. To find an eigenvector corresponding to an eigenvalue λ, we simply solve the system of linear algebraic equations (A - λI)K = 0. This system can be solved by applying the Gauss-Jordan elimination to the augmented matrix (A - λI | 0).
🔑 Definition — Eigenvalue: A number λ such that there exists a nonzero vector K satisfying AK = λK. 🔑 Definition — Eigenvector: A nonzero vector K that satisfies AK = λK for a given eigenvalue λ. 📐 Formula: (A - λI)K = 0 → This is the system to solve for eigenvectors. 📐 Formula: det(A - λI) = 0 → The characteristic equation, whose roots are the eigenvalues. 💡 Why this matters: Eigenvalues and eigenvectors are fundamental for solving systems of linear differential equations, as they allow us to find fundamental solution sets.
Example 4: Verify that the following column vector K = [1, -1, 1]ᵀ is an eigenvector of the matrix A = [[0, -1, -3], [2, 3, 3], [-2, 1, 1]]. By carrying out the multiplication AK, we see that AK = [-2, 2, -2]ᵀ = -2 * [1, -1, 1]ᵀ = (-2)K. Hence the number λ = -2 is an eigenvalue of the given matrix A.
Example 5: Find the eigenvalues and eigenvectors of A = [[1, 2, 1], [6, -1, 0], [-1, -2, -1]]. The characteristic equation is det(A - λI) = 0. Expanding with respect to the 3rd column yields -λ(λ + 4)(λ - 3) = 0, so the eigenvalues are λ = 0, -4, 3. For λ₁ = 0, Gauss-Jordan elimination gives k₁ = -(1/13)k₃, k₂ = -(6/13)k₃. Choosing k₃ = -13 yields eigenvector K₁ = [1, 6, -13]ᵀ. For λ₂ = -4, Gauss-Jordan gives k₁ = -k₃, k₂ = 2k₃. Choosing k₃ = 1 yields eigenvector K₂ = [-1, 2, 1]ᵀ. For λ₃ = 3, Gauss-Jordan gives k₁ = -k₃, k₂ = (-3/2)k₃. Choosing k₃ = -2 yields eigenvector K₃ = [2, 3, -2]ᵀ.
Note: The component k₃ could be chosen as any nonzero number. Therefore, a nonzero constant multiple of an eigenvector is also an eigenvector.
Example 6: Find the eigenvalues and eigenvectors of A = [[3, 4], [-1, 7]]. The characteristic equation is (3 - λ)(7 - λ) + 4 = 0 ⇒ (λ - 5)² = 0. Therefore, the characteristic equation has repeated real roots, and the matrix has an eigenvalue of multiplicity two: λ₁ = λ₂ = 5. To find the eigenvector(s) corresponding to λ₁ = 5, we solve (A - 5I)K = 0, which gives -2k₁ + 4k₂ = 0 and k₁ + 2k₂ = 0. It is apparent from this system that k₁ = 2k₂. Thus if we choose k₂ = 1, we find the single eigenvector K₁ = [2, 1]ᵀ.
💡 Why this matters: When a matrix has a repeated eigenvalue, it may not be possible to find a full set of linearly independent eigenvectors. This has important consequences for solving systems of differential equations.
Example 7: Find the eigenvalues and eigenvectors of A = [[9, 1, 1], [1, 9, 1], [1, 1, 9]]. The characteristic equation is (λ - 11)(λ - 8)² = 0 ⇒ λ = 11, 8, 8. Thus the eigenvalues are λ₁ = 11, λ₂ = λ₃ = 8. For λ₁ = 11, Gauss-Jordan elimination gives k₁ = k₃, k₂ = k₃. Choosing k₃ = 1 yields eigenvector K₁ = [1, 1, 1]ᵀ. For λ₂ = 8, we have (A - 8I | 0) = [[1, 1, 1, 0], [1, 1, 1, 0], [1, 1, 1, 0]]. Gauss-Jordan gives k₁ + k₂ + k₃ = 0. We are free to select two of the variables arbitrarily. Choosing k₂ = 1, k₃ = 0 yields K₂ = [-1, 1, 0]ᵀ, and choosing k₂ = 0, k₃ = 1 yields K₃ = [-1, 0, 1]ᵀ. Thus, we obtain two linearly independent eigenvectors corresponding to a single eigenvalue.
💡 Why this matters: Thus we note that when an n × n matrix A possesses n distinct eigenvalues λ₁, λ₂, ..., λₙ, a set of n linearly independent eigenvectors K₁, K₂, ..., Kₙ can be found. However, when the characteristic equation has repeated roots, it may not be possible to find n linearly independent eigenvectors of the matrix.
⭐ Key Takeaways
The eigenvalue problem reduces to finding λ such that det(A - λI) = 0, which is called the characteristic equation, and the roots of this equation are the eigenvalues. For each eigenvalue, the corresponding eigenvectors are found by solving (A - λI)K = 0 using Gauss-Jordan elimination. A nonzero constant multiple of an eigenvector is also an eigenvector. While matrices with n distinct eigenvalues always have n linearly independent eigenvectors, matrices with repeated eigenvalues may have fewer than n independent eigenvectors, which is critical for solving systems of differential equations.
🧠 Quick Revision Questions
- What is the condition for a number λ to be an eigenvalue of a matrix A?
- How do you find the eigenvectors corresponding to a given eigenvalue?
- What is the characteristic equation of a matrix and how is it used?
- How do you handle the case of repeated eigenvalues when finding eigenvectors?
- Can a matrix with repeated eigenvalues still have a full set of linearly independent eigenvectors? Give an example.