General Solution, Uniqueness, and Normalized Solutions of Second-Order ODEs

Mathematics is not about numbers, equations, computations, or algorithms: it is about understanding, William Paul Thurston

Recall

Definition. A differential equation is an equation that involves one or more dependent variables, their derivatives with respect to one or more independent variables, and the independent variables themselves, e.g., $\frac{dy}{dx} = 3x +5y, y' + y = 4xcos(2x), \frac{dy}{dx} = x^2y+y, etc.$

It involves (e.g., $\frac{dy}{dx} = 3x +5y$):

• Dependent variables: Variables that depend on one or more other variables (y).
• Independent variables: Variables upon which the dependent variables depend (x).
• Derivatives: Rates at which the dependent variables change with respect to the independent variables, $\frac{dy}{dx}$

Basic linear ODE

In mathematics, a linear ordinary differential equation (ODE) is an equation involving an unknown function and its derivatives, characterized by the property that both the function and its derivatives appear linearly. This means that the unknown function and its derivatives are raised only to the first power. They are never multiplied together, nor are they composed within nonlinear functions (such as $\sin(y)$ or $e^y$). The general form for a first-order linear ODE is: a(x)y’ + b(x)y = c(x).

Second-order Linear Homogeneous ODE’s with Constant Coefficients

In mathematics, a second-order linear homogeneous differential equation ODE with constant coefficients is an equation of the form: ay'' + by' + cy = 0 where:

• y(t) is the dependent variable, a function of the independent variable t.
• y′ and y′′ are the first and second derivatives of y with respect to t, respectively.
• t is the independent variable,
• a,b, and c are constant real numbers, with $a \ne 0$.

This equation is characterized by:

• Second-Order: The highest derivative present is the second derivative.
• Linear: y and its derivatives are all of the first power, and there are no products of y with its derivatives.
• Homogeneous: The right-hand side is zero, meaning there are no external forcing functions acting on the system.

This type of differential equation arises frequently in physics and engineering, modeling systems such as mechanical vibrations (mass-spring-damper) and electrical circuits (LRC).

General Solution of a Second-Order Linear Homogeneous ODE

When solving ordinary differential equations (ODEs), our primary goal is to find the general solution, which represents the complete family of all possible valid solutions that satisfy the equation.

For a second-order linear homogeneous ODE, the general solution is defined as: $\boxed{y(x) = c_1y_1(x) + c_2y_2(x)}$ where:

• $y_1$ and $y_2$ are two linearly independent solutions to the ODE.
• $c_1$ and $c_2$ are arbitrary constants determined by the specific initial or boundary conditions of a given problem.

Solving the Initial Value Problem and Uniqueness

Let $y_1(x)$ and $y_2(x)$ be two linearly independent solutions to a general second-order linear homogeneous differential equation of the form: $y'' + p(x)y' + q(x)y = 0$

Proposition: The family of solutions $y(x) = c_1y_1(x) + c_2y_2(x)$ is sufficient to satisfy any arbitrary initial conditions for the second-order linear homogeneous differential equation $y'' + p(x)y' + q(x)y = 0$.

Proof.

Let $y(x_0) = a$ and $y'(x_0) = b$ be the given initial values at some point $x_0$

The general solution to the homogeneous equation, constructed from our two independent solutions, is: $y(x) = c_1y_1(x) + c_2y_2(x)$ where $y_1$ and $y_2$ are two independent solutions to the differential equation.

Taking the first derivative of this general solution yields: $y'(x) = c_1y_1'(x) + c_2y_2'(x)$

We then substitute the initial conditions $y(x_0) = a$ and $y'(x_0) = b$ into these expressions to create a system of two linear equations with two unknowns ($c_1$ and $c_2$):

$$\begin{cases} c_1y_1(x_0) + c_2y_2(x_0) = a \\ c_1y_1'(x_0) + c_2y_2'(x_0) = b \end{cases}$$

From linear algebra, we know this system will have a unique solution for $c_1$ and $c_2$ if and only if the determinant of its coefficient matrix is non-zero. This determinant is called the Wronskian, evaluated at $x_0$: $W(y_1, y_2)(x_0) = \begin{vmatrix} y_1(x_0) & y_2(x_0) \\ y_1'(x_0) & y_2'(x_0) \end{vmatrix} = y_1(x_0)y_2'(x_0) - y_1'(x_0)y_2(x_0)$

This determinant plays a crucial role in the initial value problem:

• If $W(y_1, y_2)(x_0) \neq 0$, the system has exactly one unique solution for $c_1$ and $c_2$.

• If $W(y_1, y_2)(x_0) = 0$, the solutions are linearly dependent (one is a constant multiple of the other), which contradicts our initial assumption.

Conclusion: Because $y_1(x)$ and $y_2(x)$ were defined as linearly independent solutions, we are guaranteed that their Wronskian $W(y_1, y_2)(x_0) \neq 0$. Therefore, we can always solve the linear system algebraically to find specific values for $c_1$ and $c_2$. Thus, the family $\{c_1y_1(x) + c_2y_2(x)\}$ is sufficient to satisfy any valid initial conditions. $\blacksquare$

Expanding the Wronskian Definition

The Wronskian of two differentiable functions $f$ and $g$ is defined as $W(f, g) = fg' - gf'$.

More generally, for $n$ distinct functions that are $(n-1)$-times differentiable, the Wronskian is the determinant of the $n \times n$ matrix constructed by placing the functions in the first row, their first derivatives in the second row, and so on up to the $(n-1)$th derivative:

$$W(f_1, f_2, \dots, f_n) = \begin{vmatrix} f_1(x) & f_2(x) & \dots & f_n(x) \\ f_1'(x) & f_2'(x) & \dots & f_n'(x) \\ \vdots & \vdots & \ddots & \vdots \\ f_1^{(n-1)}(x) & f_2^{(n-1)}(x) & \dots & f_n^{(n-1)}(x) \end{vmatrix}$$

In our specific case for second-order ODEs ($n=2$), this matrix condenses simply to $W(y_1, y_2) = \begin{vmatrix} y_1 & y_2 \\ y_1' & y_2' \end{vmatrix} = y_1(x)y_2'(x) - y_1'(x)y_2(x)$

Theorem on the Wronskian (Abel's Identity). If $y_1$ and $y_2$ are solutions to the linear homogeneous ODE $y'' + p(x)y' + q(x)y = 0$, and the coefficient functions $p(x)$ and $q(x)$ are continuous on an interval $I$, then the Wronskian $W(y_1, y_2)$ either vanishes identically (i.e., $W(x) = 0$ for all $x \in I$) or is never zero on $I$ (i.e., $W(x) \neq 0$ for all $x \in I$).

Implications

This theorem essentially tells us that linear independence is a global property across the interval, not a local one.

• Robust Linear Independence: Because $y_1$ and $y_2$ are assumed to be linearly independent solutions, their Wronskian $W(y_1, y_2) \neq 0$ everywhere on $I$. There are no single points on the continuous interval where the solutions suddenly cross paths in a way that makes them mathematically dependent.

• Total Uniqueness of Solutions: Given any initial condition $y(x_0) = a$ and $y'(x_0) = b$ at any valid point $x_0$, the non-zero Wronskian guarantees that we can always find a unique pair of values for $c_1$ and $c_2$. Therefore, for every specific initial value problem, there exists one, and only one, valid solution curve.

Completeness of the Solution Set

Claim: The set of solutions $y(x) = c_1y_1(x) + c_2y_2(x)$ forms the complete family of solutions to the differential equation.

Note: Once you find two linearly independent solutions, $y_1$ and $y_2$, you have found the “building blocks” for all possible solutions. However, it is important to note that $y_1$ and $y_2$ are not unique. They are not the only valid basis for the solution space.

It is entirely possible to find a different pair of independent solutions, say $u_1(x)$ and $u_2(x)$, that also form a valid basis. In this scenario, the general solution is just as correctly expressed as: $y(x) = c_1u_1(x) + c_2u_2(x)$

Relationship Between Different Solution Pairs

Because the solution space is exactly two-dimensional, any valid solution can be written as a combination of any valid basis. This means the new pair ($u_1$, $u_2$) is simply a linear combination of the old pair ($y_1$, $y_2$):

$\begin{cases} u_1(x) = A_{11}y_1(x) + A_{12}y_2(x) \\ u_2(x) = A_{21}y_1(x) + A_{22}y_2(x) \end{cases}$

Where the $A$ coefficients are constants

Normalized Solutions

Among all the possible pairs of basis solutions, there is a very special, highly useful pair known as normalized solutions. We typically denote these as $Y_1(x)$ and $Y_2(x)$.

Normalized solutions are particular solutions of the ODE carefully constructed to satisfy incredibly simple, specific initial conditions at a chosen point $x_0$ (almost always $x_0 = 0$):

• First Normalized Solution ($Y_1$): Initial position is 1, initial velocity is 0. $Y_1(0) = 1, \quad Y_1'(0) = 0$
• Second Normalized Solution ($Y_2$): Initial position is 0, initial velocity is 1. $Y_2(0) = 0, \quad Y_2'(0) = 1$

Purpose of Normalized Solutions

1. Ultimate convenience. They vastly simplify the algebra required to solve specific initial value problems.
2. Direct Application: As we will see below, the arbitrary constants $c_1$ and $c_2$ literally become the initial conditions themselves. Any solution satisfying initial conditions can be directly expressed as a linear combination of $Y_1$ and $Y_2$.

Deep Dive: Examples and the Power of Uniqueness

To understand how normalized solutions function as a computational shortcut, it is best to look at two classic model equations. These examples illustrate how changing a single sign completely alters the geometry of the solution space from closed circles (trigonometric) to open curves (hyperbolic).

• Example 1. The Oscillatory System ($y'' + y = 0$)

The differential equation $y'' + y = 0$ is a second-order linear homogeneous ODE with constant coefficients, physically representing an ideal, undamped harmonic oscillator.

Step 1. Find the General Solution.
We assume a solution of the form $y(x) = e^{rx}$, leading to the characteristic equation: $r^2 + 1 = 0 \implies r = \pm i$.
Because the roots are purely imaginary ($\alpha = 0, \beta = 1$), the Euler-derived standard solutions are trigonometric:
$y_1(x) = \cos(x), \quad y_2(x) = \sin(x)$
These solutions are linearly independent and form a fundamental set of solutions for the differential equation. Any arbitrary solution is a linear combination of the two: $y(x) = c_1\cos(x) + c_2\sin(x)$

Step 2. Verify and Find the Normalized Solutions
Normalized solutions $Y_1(x)$ and $Y_2(x)$ are particular solutions tailored to specific initial conditions at a baseline point, typically $x_0 = 0$.
For $Y_1(x)$, we require $Y_1(0) = 1$ and $Y_1'(0) = 0$.
For $Y_2(x)$, we require $Y_2(0) = 0$ and $Y_2'(0) = 1$.
Let’s test our standard fundamental set at $x_0 = 0$:
$\begin{aligned} y_1(0) &= \cos(0) = 1, \quad &y_1'(0) &= -\sin(0) = 0 \\ y_2(0) &= \sin(0) = 0, \quad &y_2'(0) &= \cos(0) = 1 \end{aligned}$
By sheer mathematical convenience, the standard basis functions $\cos(x)$ and $\sin(x)$ are already naturally normalized at $x = 0$. Therefore: $Y_1(x) = \cos(x) \quad \text{and} \quad Y_2(x) = \sin(x)$

• Example 2: The Saddle System ($y'' - y = 0$)

Now, let’s flip the sign. The equation $y'' - y = 0$ represents a system governed by exponential growth and decay rather than periodic oscillation

Step 1. Find the General Solution.
The characteristic equation for this ODE is: $r^2 - 1 = 0 \implies r = \pm 1$. Since the roots are both real and distinct, our standard fundamental solutions are exponentials: $y_1(x) = e^x, \quad y_2(x) = e^{-x}$
This yields the traditional general solution and its derivative: $y(x) = c_1e^x + c_2e^{-x} \implies y'(x) = c_1e^x - c_2e^{-x}$

Step 2. Derive the Normalized Solutions.
Unlike the previous example, $e^x$ and $e^{-x}$ are not normalized at $x_0 = 0$. We must calculate the specific constants $c_1$ and $c_2$ to build $Y_1$ and $Y_2$.
Constructing $Y_1(x)$ (Setting $Y_1(0) = 1$ and $Y_1'(0) = 0$):
$\begin{cases} c_1e^0 + c_2e^0 = 1 \\ c_1e^0 - c_2e^0 = 0 \end{cases} \implies \begin{cases} c_1 + c_2 = 1 \\ c_1 - c_2 = 0 \end{cases}$
From the second equation, we find $c_1 = c_2$. Substituting this into the first equation yields $2c_1 = 1$, which means $c_1 = c_2 = \frac{1}{2}$.
This gives us: $Y_1(x) = \frac{e^x + e^{-x}}{2} = \cosh(x)$
Constructing $Y_2(x)$ (Setting $Y_2(0) = 0$ and $Y_2'(0) = 1$):
$\begin{cases} c_1e^0 + c_2e^0 = 0 \\ c_1e^0 - c_2e^0 = 1 \end{cases} \implies \begin{cases} c_1 + c_2 = 0 \\ c_1 - c_2 = 1 \end{cases}$
Adding these equations yields $2c_1 = 1 \implies c_1 = \frac{1}{2}$. It follows that $c_2 = -\frac{1}{2}$. This gives us:
$Y_2(x) = \frac{e^x - e^{-x}}{2} = \sinh(x)$.
Thus, while $\{e^x, e^{-x}\}$ is a perfectly accurate fundamental set, the hyperbolic functions $\{\cosh(x), \sinh(x)\}$ serve as the much cleaner normalized fundamental set at $x = 0$.

Advantages of Normalized Solutions

Operational Advantages of Normalized Solutions

Normalized solutions decouple the initial conditions from algebraic overhead. If you know the normalized solutions $Y_1(x)$ and $Y_2(x)$ for an ODE at a given point $x_0$, solving any arbitrary Initial Value Problem (IVP) requires zero matrix algebra.

Given the general IVP: $\begin{cases} y'' + p(x)y' + q(x)y = 0 \\ y(x_0) = y_0 \\ y'(x_0) = y'_0 \end{cases}$

The definitive, unique solution is instantly written by inspection: $\boxed{y(x) = y_0Y_1(x) + y'_0Y_2(x)}$

Why this works flawlessly?

Let’s differentiate this customized equation and evaluate it at $x_0$:

• $y(x_0) = y_0Y_1(x_0) + y'_0Y_2(x_0) = y_0(1) + y'_0(0) = y_0$
• $y'(x_0) = y_0Y'_1(x_0) + y'_0Y'_2(x_0) = y_0(0) + y'_0(1) = y'_0$

The initial targets are hit automatically because the basis functions are explicitly engineered to act as binary switches for position and velocity.

The Existence and Uniqueness Theorem: The Bedrock of Completeness

The ultimate justification for using these solution families rests on a rigorous mathematical framework:

The Existence and Uniqueness Theorem. Let the coefficient functions $p(x)$ and $q(x)$ be continuous on an open interval $I$ containing the point $x_0$. Then, for any arbitrary real constants $y_0$ and $y'_0$, there exists exactly one solution $y(x)$ on the interval $I$ that satisfies the differential equation $y'' + p(x)y' + q(x)y = 0$ along with the initial conditions $y(x_0) = y_0$ and $y'(x_0) = y'_0$.

Proving Completeness via Uniqueness

This theorem does more than guarantee that a path exists —it establishes that our linear combinations form the entire, exhaustive family of possible solutions.

1. The Vector Space Span: Suppose we have built a normalized solution family $\{c_1Y_1(x) + c_2Y_2(x)\}$ at $x_0$. As proven above, we can match any initial state $(y_0, y'_0)$ by simply setting $c_1 = y_0$ and $c_2 = y'_0$. This proves existence within our family.
2. The “Wild” Solution Test: Imagine a skeptic claims they found a completely different, highly exotic solution to the ODE called $u(x)$. This function $u(x)$ has an initial state such that $u(x_0) = u_0$ and $u'(x_0) = u'_0$.
3. The Trap: We can immediately construct a solution from our own trusted family that mimics those exact initial conditions: $v(x) = u_0Y_1(x) + u'_0Y_2(x)$
4. The Uniqueness Verdict: We now have two functions, $u(x)$ and $v(x)$, that simultaneously satisfy the exact same differential equation and the exact same initial conditions. According to the Uniqueness Theorem, it is mathematically impossible for them to diverge. They must be completely identical: $u(x) \equiv v(x) \quad \forall x \in I$

Therefore, no independent solution can ever exist outside of the fundamental linear combination family. The vector space is closed, locked at exactly two dimensions, and entirely accounted for.

Bibliography

1. NPTEL-NOC IITM, Introduction to Galois Theory.
2. Algebra, Second Edition, by Michael Artin.
3. LibreTexts, Calculus and Calculus 3e (Apex). Abstract and Geometric Algebra, Abstract Algebra: Theory and Applications (Judson).
4. Field and Galois Theory, by Patrick Morandi. Springer.
5. Michael Penn, and MathMajor.
6. Contemporary Abstract Algebra, Joseph, A. Gallian.
7. YouTube’s Andrew Misseldine: Calculus. College Algebra and Abstract Algebra.
8. MIT OpenCourseWare [18.03 Differential Equations, Spring 2006], YouTube by MIT OpenCourseWare.
9. Calculus Early Transcendentals: Differential & Multi-Variable Calculus for Social Sciences.