Introduction to Differential Equations

Most of the laws of nature are not statements about quantities — they are statements about rates of change. Newton did not tell us where a planet is; he told us how its velocity changes from moment to moment. A population does not come with a formula for its size next year; it comes with a rule for how fast it grows right now. A differential equation is the mathematical language for exactly this kind of law: an equation that links a function to its own derivatives. Learning to read, classify, and solve these equations is how you turn "here is the rule for change" into "here is what actually happens."

This page builds the foundation for the entire subject. We will pin down what a differential equation really is, learn the two labels — order and linearity — that let you predict how hard an equation will be, understand why a single equation has a whole family of solutions until you add an initial condition, and see how slope fields let you see the solutions before you ever solve anything.

Learning Objectives

• Define a differential equation and distinguish ordinary (ODE) from partial (PDE) equations.
• Classify an equation by its order and decide whether it is linear or nonlinear.
• Explain the difference between a general solution (a family) and a particular solution fixed by an initial condition.
• Verify whether a candidate function actually solves a given differential equation.
• Read a slope field and sketch the solution curve through a given point.
• Connect the subject to its historical birth in Newtonian and Leibnizian mechanics.

Quick Answer

A differential equation is an equation that relates an unknown function to one or more of its derivatives, for example $\frac{dy}{dx} = 2y$. Its order is the highest derivative that appears; it is linear if the unknown function and its derivatives appear only to the first power and are not multiplied together. A solution is a function that makes the equation true when substituted in. A single equation has infinitely many solutions — a general solution with an arbitrary constant — and an initial condition like $y(0) = 3$ selects one particular solution. A slope field draws the slope $\frac{dy}{dx}$ as a short line segment at many points, so solution curves are the curves that flow along those segments.

Where It Came From

Differential equations were born in the 1660s–1680s alongside calculus itself, and for a very concrete reason: Isaac Newton wanted to explain the motion of the planets and falling bodies, and no existing mathematics could do it.

The problem was this. Everyone could see that objects move, but the tools of the day described finished quantities — a distance traveled, a fixed speed. Real motion is different: a falling apple speeds up continuously, and a planet's velocity swings around its orbit. What Newton grasped is that the fundamental laws are about instantaneous change. His second law, force equals mass times acceleration, is really the statement that acceleration — the second derivative of position — is determined by the forces acting at each instant. That is a differential equation. To even write it down, Newton had to invent the derivative (his "fluxions"), and to extract actual orbits from it he had to invent integration to undo those derivatives.

Working independently in Germany, Gottfried Wilhelm Leibniz developed the same calculus in the 1670s–1680s, and it is his notation — $\frac{dy}{dx}$ and the integral sign $\int$ — that we still use, precisely because it makes differential equations easy to manipulate. The term aequatio differentialis is Leibniz's.

Over the next century the Bernoulli family, and above all Leonhard Euler in the 1700s, turned this into a systematic subject, and Joseph-Louis Lagrange and others applied it to mechanics, fluids, and heat. The through-line never changed: whenever a scientist could state how a quantity changes, a differential equation was the bridge from that local rule to global behavior. That is still exactly why we study them.

What a Differential Equation Is (and Ordinary vs. Partial)

A differential equation is any equation containing derivatives of an unknown function. The unknown is not a number — it is a function, and solving means finding that function.

Compare an ordinary algebraic equation with a differential one:

• Algebraic: $x^2 - 5x + 6 = 0$. The unknown is a number; the solutions are $x = 2$ and $x = 3$.
• Differential: $\frac{dy}{dx} = 3x^2$. The unknown is a function $y(x)$; a solution is $y = x^3 + C$.

Notice the differential equation says: "the rate of change of $y$ equals $3x^2$." Solving it means asking "what function has that derivative?" — which is integration. Indeed $\int 3x^2 \, dx = x^3 + C$, and the arbitrary constant $C$ is our first hint that solutions come in families.

We split differential equations into two big families by how many variables the unknown function depends on:

• An ordinary differential equation (ODE) involves derivatives with respect to a single variable — e.g. $\frac{dy}{dx} = 2y$, where $y$ depends only on $x$.
• A partial differential equation (PDE) involves partial derivatives with respect to several variables — e.g. the heat equation $\frac{\partial u}{\partial t} = k\,\frac{\partial^2 u}{\partial x^2}$, where temperature $u$ depends on both position and time.

This introductory course is about ODEs. PDEs are a large and deeper subject built on the same ideas.

Worked example — is it a solution? Show that $y = e^{2x}$ solves $\frac{dy}{dx} = 2y$.

Differentiate the candidate: $\frac{dy}{dx} = 2e^{2x}$. Now substitute into the equation. The left side is $2e^{2x}$; the right side is $2y = 2e^{2x}$. They match for every $x$, so $y = e^{2x}$ is indeed a solution. (In fact so is $y = Ce^{2x}$ for any constant $C$ — check it the same way.)

Order and Linearity: The Two Labels That Predict Difficulty

Before solving any equation, classify it. Two labels tell you almost everything about which methods apply.

Order is the highest derivative that appears.

• $\frac{dy}{dx} = 2y$ is first order (highest derivative is the first).
• $\frac{d^2y}{dx^2} + 3\frac{dy}{dx} + 2y = 0$ is second order (a second derivative appears).
• $\frac{d^3y}{dx^3} = x$ is third order.

Newton's second law is second order, because acceleration is a second derivative of position — that is why so much of classical physics is second-order.

Linearity is subtler and more important. An ODE is linear if the unknown function $y$ and all its derivatives appear only to the first power, are not multiplied by each other, and sit inside no nonlinear function (no $\sin y$, $y^2$, $e^y$, $\sqrt{y}$, etc.). The coefficients may depend on $x$ freely. A general linear ODE looks like:

$a_n(x)\,y^{(n)} + \cdots + a_1(x)\,y' + a_0(x)\,y = g(x)$

Anything that violates this is nonlinear. Let us classify a few:

Equation	Order	Linear?	Reason
$y' + 3y = x$	1	Linear	$y$, $y'$ to first power
$y'' + 5y' + 6y = 0$	2	Linear	first powers only
$y' = y^2$	1	Nonlinear	$y^2$ is second power
$y'' + \sin y = 0$	2	Nonlinear	$\sin y$ is nonlinear in $y$
$x^2 y' + y = \cos x$	1	Linear	coefficient depends on $x$, that's fine
$y \cdot y' = 1$	1	Nonlinear	$y$ multiplies $y'$

Why care? Linear equations are vastly more tractable. They obey the superposition principle (sums of solutions are solutions), have well-developed general theory, and many have neat closed-form solutions. Nonlinear equations — like the pendulum's $y'' + \sin y = 0$ — often have no elementary formula and must be studied qualitatively or numerically. So the moment you see $\sin y$ multiplying or a $y^2$, brace yourself.

Solutions, General vs. Particular, and Initial Conditions

Here is the single most important structural fact for beginners: a differential equation does not have one solution — it has an infinite family, until you pin it down.

Consider $\frac{dy}{dx} = 2x$. Integrating, $y = x^2 + C$. Every value of $C$ gives a valid solution — the parabolas $y = x^2$, $y = x^2 + 1$, $y = x^2 - 5$ all have slope $2x$ everywhere. This family, complete with its arbitrary constant, is the general solution.

To choose one curve you supply extra data. An initial condition specifies the function's value at a point, such as $y(0) = 4$. Combined with the general solution this gives an initial value problem (IVP).

Worked example — solving an IVP. Solve $\frac{dy}{dx} = 2x$ with $y(0) = 4$.

1. General solution: $y = x^2 + C$.
2. Apply the condition: at $x = 0$, $y = 4$, so $4 = 0^2 + C$, giving $C = 4$.
3. Particular solution: $y = x^2 + 4$.

A rule of thumb: an $n$th-order equation needs $n$ conditions to fix a unique solution, because integrating $n$ times introduces $n$ arbitrary constants. A second-order equation of motion, for instance, needs both an initial position and an initial velocity — which matches physical intuition: to predict a thrown ball's path you must know where it started and how fast it left your hand.

Worked example — a second-order family. The equation $\frac{d^2y}{dx^2} = -y$ has general solution $y = A\cos x + B\sin x$, with two constants. Verify: $y' = -A\sin x + B\cos x$, then $y'' = -A\cos x - B\sin x = -(A\cos x + B\sin x) = -y$. It checks. Fixing $y(0)$ and $y'(0)$ would determine both $A$ and $B$.

Slope Fields: Seeing Solutions Without Solving

Many equations cannot be solved with a formula, yet we can still see their solutions. A first-order equation $\frac{dy}{dx} = f(x, y)$ tells you the slope of the solution curve at every point $(x, y)$ in the plane. If at each of many points you draw a short line segment with that slope, you get a slope field (or direction field). Solution curves are simply the curves that stay tangent to these segments everywhere — like iron filings tracing a magnetic field.

Worked example — building a slope field for $\frac{dy}{dx} = x$. Here the slope depends only on $x$:

Point $(x, y)$	Slope $= x$	Segment tilts
$(-2, \text{any})$	$-2$	steeply down
$(-1, \text{any})$	$-1$	down at 45°
$(0, \text{any})$	$0$	flat
$(1, \text{any})$	$1$	up at 45°
$(2, \text{any})$	$2$	steeply up

Every point in a vertical column has the same slope. Following the flow, the curves that hug this field are the parabolas $y = \tfrac{1}{2}x^2 + C$ — which is exactly the general solution you'd get by integrating $\frac{dy}{dx} = x$. Starting from a chosen point, say $(0, 1)$, you trace the one parabola through it: $y = \tfrac{1}{2}x^2 + 1$.

The power of slope fields is qualitative insight. From the field of $\frac{dy}{dx} = 2 - y$ you can see that wherever $y < 2$ the slope is positive (solutions rise) and wherever $y > 2$ the slope is negative (solutions fall), so all solutions drift toward the horizontal line $y = 2$ — an equilibrium — without solving a thing.

Real-World Applications

• Physics and engineering: Newton's second law $m\frac{d^2x}{dt^2} = F$ governs everything from projectile motion to spacecraft trajectories. Circuit equations for current and voltage, and the vibration of bridges and buildings, are differential equations.
• Biology and medicine: Population growth ($\frac{dP}{dt} = kP$ for exponential, or the logistic model for limited resources), the spread of epidemics (SIR models), and drug concentration decay in the bloodstream (pharmacokinetics) are all ODEs.
• Economics and finance: Models of continuous compound interest ($\frac{dA}{dt} = rA$), and the option-pricing Black–Scholes equation (a PDE), rest on differential equations.
• Chemistry: Reaction-rate laws express how concentrations change over time as differential equations.
• Climate and weather: Forecasting solves enormous systems of differential equations describing heat, moisture, and airflow.

Common Mistakes

Mistake 1: Thinking the solution is a number. Students fresh from algebra expect an answer like $x = 3$. But the unknown in a differential equation is a function, so the answer is something like $y = e^{2x}$ or $y = x^2 + C$. Correction: always ask "what function makes this true?", not "what number?".

Mistake 2: Forgetting the arbitrary constant (or that there are several). Writing $y = x^2$ instead of $y = x^2 + C$ throws away infinitely many solutions and makes it impossible to satisfy an initial condition. Correction: every integration introduces a constant; a first-order equation carries one, an $n$th-order equation carries $n$. Only after applying the initial conditions do the constants get pinned down.

Mistake 3: Confusing order with degree, or misjudging linearity. Seeing $\left(\frac{dy}{dx}\right)^2$ and calling the equation "second order" is wrong — it is still first order (highest derivative is the first), but it is nonlinear because that derivative is squared. Correction: order = highest derivative present; linearity = whether $y$ and its derivatives appear only to the first power and unmultiplied. They are independent labels.

Comparison and Connections

Differential equations sit right on top of calculus. Solving the simplest ones is integration; the difference is that once $y$ appears on the right-hand side (as in $\frac{dy}{dx} = 2y$), you can no longer just integrate — you need dedicated techniques.

Idea	What the unknown is	What "solving" means
Algebraic equation	a number	find the value(s) of $x$
Integral $\int f\,dx$	a function	antidifferentiate $f$
Differential equation	a function	find $y$ from a relation among $y$ and its derivatives

It is also worth distinguishing the two ways to attack an equation. An analytic (closed-form) solution is an exact formula, available mainly for linear and a few special nonlinear equations. A qualitative or numerical solution — slope fields, equilibrium analysis, or step-by-step methods like Euler's method — describes or approximates behavior when no formula exists. Real scientific practice uses both.

Practice Questions

Recall

State the order and linearity of each: (a) $y'' + 4y = 0$, (b) $y' = y^3$, (c) $x y' + y = e^x$.

Answer: (a) second order, linear; (b) first order, nonlinear (the $y^3$); (c) first order, linear (the $x$ coefficient is allowed).

Understanding

Explain in your own words why $\frac{dy}{dx} = 2x$ has infinitely many solutions but $\frac{dy}{dx} = 2x,\ y(1) = 5$ has exactly one.

Guidance: Integrating gives $y = x^2 + C$, one curve for each $C$ — a whole family of parabolas. The condition $y(1) = 5$ forces $5 = 1 + C$, so $C = 4$, selecting the single curve $y = x^2 + 4$.

Application

Verify that $y = 3e^{-x}$ solves $\frac{dy}{dx} + y = 0$, then find the particular solution of $\frac{dy}{dx} + y = 0$ satisfying $y(0) = 7$.

Answer: $\frac{dy}{dx} = -3e^{-x}$, and $\frac{dy}{dx} + y = -3e^{-x} + 3e^{-x} = 0$. ✓ The general solution is $y = Ce^{-x}$; applying $y(0) = 7$ gives $C = 7$, so $y = 7e^{-x}$.

Analysis

Look at the equation $\frac{dy}{dx} = y(1 - y)$ without solving it. What are its equilibrium solutions, and toward which one do solutions starting at $0 < y < 1$ move?

Guidance: The slope is zero when $y = 0$ or $y = 1$ — the equilibria. For $0 < y < 1$ both factors are positive, so $\frac{dy}{dx} > 0$ and solutions rise toward $y = 1$. This is the logistic model, and $y = 1$ is the stable carrying capacity.

FAQ

Is solving a differential equation just integration? For the simplest form $\frac{dy}{dx} = f(x)$, yes — you integrate. But as soon as the right side involves $y$ itself, plain integration fails and you need methods like separation of variables, integrating factors, or characteristic equations. Integration is one tool in a much larger kit.

Why does one equation have infinitely many solutions? That feels wrong. Because the equation only constrains the rate of change, not the starting value. Countless curves can share the same slope pattern while sitting at different heights — think of a family of parallel parabolas. The initial condition supplies the missing starting information and picks out one curve.

What's the practical difference between linear and nonlinear equations? Linear equations have a complete, reliable theory and often exact solutions; you can add solutions together to build new ones. Nonlinear equations frequently have no formula at all and can behave chaotically, so we study them with slope fields, stability analysis, and computers. Recognizing which you have tells you what to even attempt.

Do I always need an initial condition? Only if you want a unique solution. The general solution (with its arbitrary constants) fully answers "what are all the solutions?" You add initial (or boundary) conditions when a specific real situation fixes the starting state.

What's a slope field good for if I can just solve the equation? Two things. First, many equations cannot be solved with a formula, and the slope field still reveals the behavior. Second, even for solvable equations it builds intuition — you can see at a glance where solutions rise, fall, or level off at equilibria, which a formula alone may hide.

Quick Revision

• A differential equation relates an unknown function to its derivatives; the answer is a function, not a number.
• ODE: derivatives in one variable. PDE: partial derivatives in several. This course is ODEs.
• Order = highest derivative present. Linear = $y$ and derivatives appear only to first power, unmultiplied, inside no nonlinear function; coefficients may depend on $x$.
• General solution = family with arbitrary constant(s); particular solution = one member fixed by initial conditions. An $n$th-order equation needs $n$ conditions.
• Verify a solution by substituting it and checking both sides match.
• Slope field: draw slope $\frac{dy}{dx} = f(x,y)$ as segments; solution curves flow tangent to them; flat segments mark equilibria.
• Key example: $\frac{dy}{dx} = ky \Rightarrow y = Ce^{kx}$ (exponential growth/decay).

Related Topics

Prerequisites

• Differential Equations overview
• Differentiation and the meaning of the derivative (Calculus)
• Integration and antiderivatives (Calculus)

Related Topics

• Slope fields and qualitative behavior of solutions
• Linear versus nonlinear systems

Next Topics

• First-order separable equations
• First-order linear equations and integrating factors
• Second-order linear differential equations