Chapter 3 – High-Order Differential Equations

3.1 Theory of Linear Equations

3.1.1 What is an IVP?

For a linear differential equation, an n^{th}-order Initial-Value Problem (IVP) is

(1)   \begin{equation*}a_n (x) \frac{d^n y}{dx^n} + a_{n-1} (x) \frac{d^{n-1}y}{dx^{n-1}} + … + a_1 (x) \frac{dy}{dx} = g(x),\end{equation*}


which is solved using the given Initial-Value Conditions (IVC) as

(2)   \begin{equation*}y(x_0) = y_0, \quad y^\prime (x_0) = y_1, \quad …, \quad y^{(n-1)} (x_0) = y_{n-1} .\end{equation*}

3.1.2 What is a BVP?

For a linear differential equation, a second-order Boundary-Value Problem (BVP) is

(3)   \begin{equation*}a_2(x) \frac{d^2y}{dx^2} + a_1 (x) \frac{dy}{dx} + a_0 (x) y = g(x) ,\end{equation*}


which is solved using the Boundary Condition (BC) given at some points as

(4)   \begin{equation*}y(a) = y_0, \quad y(b) = y_1 .\end{equation*}

3.1.3 Linearly Dependent Functions

A set of functions are linearly dependent in an interval I, if constants of c_1, c_2, …, c_n exist in such a way that

(5)   \begin{equation*}c_1 f_1 (x) + c_2 f_2 (x) + … + c_n f_n (x) = 0 ,\end{equation*}


for every x in the interval I. Note that all the constants cannot be zero.

3.1.4 Criterion for Linearly Independent Solutions

A set of solutions of the homogeneous linear n^{th}-order DE on an interval I are linearly independent if and only if the Wronskian of the solutions is non-zero.

3.1.5 Wronskian of a Set of Functions

The Wronskian of a set of functions is a determinant of a matrix and defined as follows, provided that all the functions have at least n-1 derivatives,

(6)   \begin{equation*}W(f_1, f_2, \dots, f_n) = \begin{vmatrix} f_1 & f_2 & \dots & f_n \\ f^\prime_1 & f^\prime_2 & \dots & f^\prime_n \\ \vdots & \vdots & \vdots & \vdots \\ f^{(n)}_1 & f^{(n)}_2 & \dots & f^{(n)}_n \end{vmatrix}\end{equation*}

3.2 Reduction of Order

3.2.1 What is Reduction of Order method?

Suppose y_1 (x) is a solution to the following homogeneous linear second-order differential equation.

(7)   \begin{equation*}a_2(x) y^{\prime \prime} + a_1 (x) y^\prime + a_0 (x) y = 0.\end{equation*}


Recall that the solutions, y_1(x) and y_2(x), are linearly independent on an interval I. Thus, their ratio \frac{y_2(x)}{y_1(x)} is non-constant on the same interval, meaning that \frac{y_2(x)}{y_1(x)} = u(x).
Now, by solving the first-order differential equation of y_2(x) = u(x) y_1(x), we can find the second solution as well.

3.2.2 How does Reduction of Order work?

By dividing the homogeneous linear second-order differential equation by a_2(x), we will recover the standard form as

(8)   \begin{equation*}y^{\prime \prime} + P(x) y^\prime + Q(x) y = 0\end{equation*}


Suppose that y_1(x) is a known solution of the DE and y_1(x) \neq 0, then

(9)   \begin{equation*}y_2(x) = y_1(x) \int \frac{e^{- \int P(x) dx}}{y_1^2 (x)} dx .\end{equation*}

3.3 Homogeneous Linear Differential Equations with Constant Coefficients

3.3.1 Form an Auxiliary Equation

Consider the following second-order homogeneous linear differential equation with constant coefficients

(10)   \begin{equation*}a y^{\prime \prime} + b y^\prime + c y = 0.\end{equation*}


Assume that the solution has the general form of y = e^{mx}, then after substituting y^\prime and y^{\prime \prime} into the equation, the auxiliary equation will be found as

(11)   \begin{equation*}a m^2 + bm + c = 0\end{equation*}


This equation will have the following two roots

(12)   \begin{equation*}m_{1,2} = \frac{-b \pm \sqrt{b^2 - 4ac}}{2a}\end{equation*}

3.3.2 Different Scenarios for the Auxiliary Equation

  1. \sqrt{b^2 - 4ac} > 0 ~: \quad m_1 and m_2 are real and distinct \quad \rightarrow \quad y = c_1 e^{m_1 x} + c_2 e^{m_2 x}.
  2. \sqrt{b^2 - 4ac} = 0 ~: \quad m_1 and m_2 are real and equal \quad \rightarrow \quad y = c_1 e^{m_1 x} + c_2 x e^{m_1 x}.
  3. \sqrt{b^2 - 4ac} < 0 ~: \quad m_1 and m_2 are conjugate complex numbers, and m_1 = \alpha + i \beta and m_2 = \alpha - i \beta. Then, y = e^{\alpha x} \left( c_1 \cos{\beta x} + c_2 \sin{\beta x} \right).

3.3.3 Higher-Order Equations

Consider the following high-order homogeneous linear differential equation with constant coefficients

(13)   \begin{equation*}a_n y^{(n)} + a_{n-1} y^{(n-1)} + \dots + a_1 y^{prime} + a_0 y = 0,\end{equation*}


the auxiliary equation will be an n-th degree polynomial

(14)   \begin{equation*}a_n m^n + a_{n-1} m^{n-1} + \dots + a_1 m + a_0 = 0.\end{equation*}


The solution to the differential equation depends on the roots of the auxiliary equation and is determined similarly to the previous section.

3.4 Non-Homogeneous Linear Differential Equations

A non-homogeneous linear differential equation has a general form of

(15)   \begin{equation*}a_n y^{(n)} + a_{n-1} y^{(n-1)} + ... + a_1 y^\prime + a_0 y = g(x).\end{equation*}


We need to find the complementary function, y_c, and also any particular solution, y_p, of the non-homogeneous equation. Then, the solution to the given differential equation will be y=y_c + y_p. The complementary function, y_c, is the solution to the associated homogenous differential equation, which is

(16)   \begin{equation*}a_n y^{(n)} + a_{n-1} y^{(n-1)} + ... + a_1 y^\prime + a_0 y = 0.\end{equation*}


The complementary function is found using the methods that we learnt so far; however, there are different methods to find the particular solutions.

3.4.1 Method of Undetermined Coefficients

The idea of this method is an educated guess about the form of y_p based on the input function g(x). This method is only suitable if the coefficients, a_i, are constants, and g(x) is a constant, a polynomial function, an exponential function e^{\alpha x}, sine or cosine functions \sin \beta x or \cos \beta x, or finite sums and products of these functions.
The following functions are some examples of the types of inputs g(x) that are appropriate:

    \begin{equation*}g(x) = 10, \end{equation*}


    \begin{equation*}g(x) = x^2 - 5x,\end{equation*}


    \begin{equation*}g(x) = 15x - 6 + 8e^{-x},\end{equation*}


    \begin{equation*}g(x) = \sin 3x - 5 x \cos 2x,\end{equation*}


    \begin{equation*}g(x) = x e^x \sin x + (3 x^2 - 1) e^{-4x}.\end{equation*}


And here are some examples of the types of inputs g(x) that are not appropriate:

    \begin{equation*}g(x) = ln x\end{equation*}


    \begin{equation*}g(x) = \frac{1}{x}\end{equation*}


    \begin{equation*}g(x) = \tan x\end{equation*}


    \begin{equation*}g(x) = \sin ^{-1} x\end{equation*}

To find a particular solution to a non-homogeneous linear differential equation using the method of undetermined coefficients, we first need to assume a form of y_p based on g(x) and then substitute y_p in the differential equation to find the coefficients.

g(x) = 1y_p = A
g(x) = 5x + 7y_p = Ax+B
g(x) = 10x^2 - 8xy_p = Ax^2 + Bx + C
g(x) = x^3 - 2x^2y_p = Ax^3 + Bx^2 + Cx + D
g(x) = \sin 4xy_p = A \cos 4x + B \sin 4x
g(x) = \cos 2xy_p = A \cos 4x + B \sin 4x
g(x) = e^{5x}y_p = A e^{5x}
g(x) = (x-1) e^{5x}y_p = (Ax+B) e^{5x}
g(x) = 2x^2 e^{5x}y_p = (Ax^2 +Bx + C) e^{5x}
g(x) = e^{3x} \sin 4xy_p = Ae^{3x} \cos 4x + B e^{3x} \sin 4x
g(x) = 5x^2 \sin 4xy_p = (Ax^2+Bx+C) \cos 4x + (Ex^2+Fx+G) \sin 4x
g(x) = xe^{3x} \cos 4xy_p = (Ax+B) \cos 4x + (Cx+D) e^{3x} \sin 4x

Note that if any y_{p_{i}} contains terms that duplicate terms in y_c, then that y_{p_{i}} must be multiplied by x^n, where n is the smallest positive integer that eliminates that duplication.

3.4.2 Method of Variation of Parameters

First, we need to write the linear second-order differential equation in the standard form,

(17)   \begin{equation*}y^{\prime \prime} + P(x) y^{\prime} + Q(x) y = f(x).\end{equation*}


Then, the particular solution is y_p = u_1 (x) y_1 (x) + u_2 (x) y_2 (x), where y_1 and y_2 form a fundamental set of solutions of the associated homogenous form of the differential equation, and

(18)   \begin{equation*}u^{\prime}_1 = \frac{W_1}{W} = - \frac{y_2 f(x)}{W} \rightarrow u_1(x) = - \int_{x_0}^x \frac{y_2(t) f(t)}{W(t)}\end{equation*}


(19)   \begin{equation*}u^{\prime}_2 = \frac{W_2}{W} = \frac{y_1 f(x)}{W} \rightarrow u_2(x) = \int_{x_0}^x \frac{y_1(t) f(t)}{W(t)}\end{equation*}


And, for higher-order non-homogeneous differential equations of the form

(20)   \begin{equation*}y^{(n)} + P_{n-1} (x) y^{(n-1)} + ... + P_1(x) y^\prime + P_0(x) y = f(x),\end{equation*}


we have

(21)   \begin{equation*}u^{\prime}_k = \frac{W_k}{W}, \quad k=1,2,...,n.\end{equation*}

3.5 Cauchy-Euler Differential Equations

The Cauchy-Euler equation, Euler-Cauchy equation, Euler equation, or equidimensional equation, is any linear differential equation of the form

(22)   \begin{equation*}a_n x^n \frac{d^n y}{dx^n} + a_{n-1} x^{n-1} \frac{d^{n-1}y}{dx^{n-1}} + ... + a_1 x \frac{dy}{dx} + a_0 y = g(x), \end{equation*}


where the coefficients a_n, a_{n-1}, ..., a_0 are constants.
To find the complementary function, i.e. the solution to the associated homogeneous differential equation, we assume that the solution has a general form of y = x^m, and find its derivatives. Then, we substitute the obtained derivates inside the differential equation to find the auxiliary equation.

3.5.1 Second-Order Cauchy-Euler Differential Equation

The homogeneous second-order linear Cauchy-Euler differential equation has a general form of

(23)   \begin{equation*}ax^2 \frac{d^2 y}{dx^2} + bx \frac{dy}{dx} + cy = 0.\end{equation*}


The auxiliary equation is

(24)   \begin{equation*}am^2 + (b-a) m + c = 0.\end{equation*}

Based on the roots of the auxiliary equation, there will be three different solutions to the given ordinary differential equation.

  1. Distinct Real Roots. y = c_1 x^{m_1} + c_2 x^{m_2}.
  2. Repeated Real Roots. y = c_1 x^{m_1} + c_2 x^{m_1} ln x.
  3. Conjugate Complex Roots. y = x^{\alpha} \left( c_1 \cos \left( \beta ln x \right) + c_2 \sin \left( \beta ln x \right) \right).

3.6 High-Order Linear Differential Equations with Variable Coefficients

3.6.1 Power Series Solution

Whenever the coefficients of a high-order differential equation are variable, we assume that the solution has a form of power series. The solution and its derivatives will have the following form

(25)   \begin{equation*}y = \sum_{n=0}^{\infty} c_n x^n,\end{equation*}


(26)   \begin{equation*}y^{\prime} = \sum_{n=1}^{\infty} n c_n x^{n-1},\end{equation*}


(27)   \begin{equation*}y^{\prime \prime} = \sum_{n=2}^{\infty} n (n-1) c_n x^{n-2},\end{equation*}


    \begin{equation*}\vdots\end{equation*}


Then, we will substitute the above derivatives in the original differential equation and shift the summation index so that all of the power series are of degree n. By finding the constants of c_k, the solutions to the given differential equations are found.

3.7 Problems

Reference

Dennis G. Zill. Advanced Engineering Mathematics, 6^{th} edition. Jones & Bartlett Learning. 2016.