Substitution Methods

The Art of Transformation

By now, you have several techniques for solving first-order differential equations: separation of variables, integrating factors for linear equations, and the potential function method for exact equations. But many equations stubbornly refuse to fit any of these forms.

The key insight of this chapter is that an equation that looks unsolvable might actually be a solvable equation in disguise. The right substitution can transform a complicated problem into one you already know how to solve.

This is pattern recognition at its finest. Once you learn to spot certain structures, equations that seemed impossibly tangled become straightforward.

Homogeneous Equations

The word "homogeneous" means different things in different contexts. In linear algebra, it refers to equations equal to zero. Here, it means something quite specific: a first-order equation is homogeneous if it can be written in the form

where the right-hand side depends only on the ratio $y/x$, not on $x$ and $y$ separately.

How do you recognize such equations? Look for terms where $x$ and $y$ appear with the same total degree. For instance, $\frac{x^2 + y^2}{xy}$ is homogeneous because the numerator has degree 2 and the denominator has degree 2, making the overall degree 0. You can always factor out powers of $x$ and be left with a function of $y/x$ alone.

The Substitution v = y/x

Once you recognize a homogeneous equation, the substitution that cracks it open is

This means $y = vx$, where both $v$ and $x$ are variables. To find $dy/dx$, we use the product rule:

This is the crucial transformation. The original equation $dy/dx = f(y/x)$ becomes

Rearranging:

This is always separable:

What seemed like a complicated equation involving both $x$ and $y$ has become a separable equation in $v$ and $x$. After integrating both sides, we substitute back $v = y/x$ to get the solution in terms of $y$ and $x$.

Bernoulli Equations

Another important class of equations that yield to substitution are Bernoulli equations, which have the form

where $n$ is any real number except 0 or 1. When $n = 0$, the equation is linear. When $n = 1$, the $y^n$ term combines with the $P(x)y$ term, again giving a linear equation. But for other values of $n$, the $y^n$ term makes the equation nonlinear.

The remarkable fact is that the substitution

always transforms a Bernoulli equation into a linear equation in $v$.

To see why this works, start by dividing the original equation by $y^n$:

Now notice that $\frac{dv}{dx} = (1-n)y^{-n}\frac{dy}{dx}$, so $y^{-n}\frac{dy}{dx} = \frac{1}{1-n}\frac{dv}{dx}$. Substituting:

Multiplying through by $(1-n)$:

This is a standard linear first-order equation in $v$, which you can solve using integrating factors. Once you find $v$, you recover $y = v^{1/(1-n)}$.

Why These Substitutions Work

The substitutions for homogeneous and Bernoulli equations are not arbitrary tricks. They work because they exploit the structure of the equation.

For homogeneous equations, the key is that $f(y/x)$ treats $y$ and $x$ as a ratio. By making that ratio into a new variable $v$, we capture the essential structure and eliminate the interaction between $x$ and $y$.

For Bernoulli equations, the $y^n$ term creates nonlinearity. The substitution $v = y^{1-n}$ is precisely chosen so that when you differentiate and substitute, the powers of $y$ cancel in exactly the right way to leave a linear equation.

This is a general principle in mathematics: when an equation has hidden structure, the right change of variables can make that structure explicit and turn a hard problem into an easy one.

Comparing Approaches

To appreciate the power of substitution, consider what happens when you try to solve a homogeneous equation without recognizing its structure.

The direct approach—trying to separate variables, checking for exactness—leads to dead ends or complicated integrating factors. The substitution approach cuts straight through to the solution.

Building Intuition

Recognizing when to use a substitution comes with practice. Here are some heuristics:

Check for homogeneity: If every term in your equation has the same degree (counting $x$ and $y$ together), try $v = y/x$. Expressions like $\frac{y}{x}$, $\frac{x + y}{x - y}$, or $\frac{x^2 + xy}{y^2}$ are clues.

Spot Bernoulli form: If your equation is "almost linear" except for a $y^n$ term with $n \neq 0, 1$, use the Bernoulli substitution $v = y^{1-n}$.

Consider the goal: A good substitution simplifies the equation's structure. Ask yourself: what makes this equation hard? Can I introduce a new variable that absorbs that complexity?

Sometimes you will try a substitution and it will not work cleanly. That is information too—it tells you to look for a different structure.

Beyond These Methods

Homogeneous and Bernoulli equations are the most common cases where substitution helps, but the principle extends further. Some equations become tractable with substitutions like $v = x + y$, $v = xy$, or $v = y^2$. The art is in recognizing what transformation will expose the underlying solvable structure.

In later chapters, you will see even more powerful transformations: the Laplace transform, which converts differential equations into algebraic ones, and series substitutions that handle equations with no closed-form solution. The theme remains the same—transformation is one of the most powerful tools in the differential equations toolkit.

Key Takeaways

• Substitution transforms difficult equations into forms you already know how to solve
• Homogeneous equations have the form $dy/dx = f(y/x)$; use $v = y/x$, so $dy/dx = v + x(dv/dx)$
• Bernoulli equations have the form $y' + P(x)y = Q(x)y^n$; use $v = y^{1-n}$ to linearize
• The chain rule is essential: always compute how $dy/dx$ transforms under your substitution
• Pattern recognition improves with practice—learn to spot homogeneous and Bernoulli structures
• When stuck, ask: what substitution would simplify this equation's structure?