Chapter 9: Differential Equations

Welcome to Last Minute Lecture.

This free chapter overview is designed to help students review and understand key concepts.

These summaries supplement not replaced the original textbook and may not be redistributed or resold.

For complete coverage, always consult the official text.

Welcome curious learners to another deep dive.

Today we're diving into a foundational concept in mathematics, one that truly brings calculus to life.

Differential equations.

From the ebb and flow of animal populations to, I don't know, the intricate workings of a circuit, these mathematical models are kind of the unsung heroes of understanding how things change.

Indeed.

Whenever scientists, whether in physics, biology, or even economics, use calculus to model a dynamic phenomenon,

they're almost certainly analyzing a differential equation.

Right.

It's not just about finding a value at a point, it's really about predicting how systems evolve over time based on their current rate of change.

That deeper understanding of change is truly what makes them so powerful.

Our mission today is to unpack the core ideas behind differential equations.

We're drawing insights from chapter nine of calculus, early transcendentals by Stuart, Clegg, and Watson.

A great resource.

Think of this as your essential guide to understanding a field that models everything from, say, the thickness of sea ice over decades to the subtle currents in your electronics.

We'll demystify what they are, how they arise, and the various powerful ways we can either solve them or gain deep insights into their behavior.

Sounds good.

So to kick things off, for someone encountering this for the first time, what's the fundamental idea behind a differential equation?

Well, at its simplest, a differential equation is an equation that connects a function to its rates of change, its derivatives.

So instead of just saying y equals something, we're saying the rate of change of y equals something.

It's how we mathematically represent situations where the future behavior of a system is influenced by its current state and how it's changing right now.

That makes perfect sense.

And for many of us, our first encounter with them often comes through a classic example.

Population growth.

How does that work?

Yeah, that's an excellent starting point.

Imagine a population, say, bacteria in a petri dish with unlimited resources.

A very simple model suggests that the population grows at a rate directly proportional to its current size.

So more bacteria means more new bacteria.

Exactly.

More means more.

And what's fascinating is that this simple assumption, when you put it into a differential equation, leads directly to exponential growth.

Exponential growth.

That sounds like it could get out of hand quickly.

Precisely.

And this model, while foundational, it quickly reveals a key insight.

Real world limits must be considered.

Right.

Left unchecked, it predicts infinite growth, which, you know, we know isn't true for any biological system.

It's a mathematical curiosity that immediately shows us where its practical limitations lie.

So if infinite resources aren't real, how do we make the model more, well, realistic?

That's where the logistic differential equation comes in.

It introduces a crucial concept.

Carrying capacity.

Carrying capacity.

Which is basically the maximum population an environment can sustainably support.

This additional factor in the equation means that as the population gets closer to this limit, m, its growth rate naturally slows down.

And what's truly insightful here is how this single equation captures a complete story.

Almost like a narrative.

Absolutely.

If the population, p, is small compared to m, the equation behaves much like the simpler exponential growth rapid increase.

Okay.

But as p approaches that carrying capacity, m, the growth rate dramatically tapers off.

It slows right down.

And if it goes over?

Good question.

If, by some chance, the population overshoots the capacity, the model even predicts it will decrease back towards that stable equilibrium point at m, it shows us how systems can naturally regulate themselves.

It finds a balance.

Fascinating.

Beyond living populations, I know differential equations pop up everywhere in the physical world.

A classic physics example is a simple spring, right?

You're right.

The applications are incredibly broad.

Think about a weight bouncing on a spring.

Hooke's law tells us the restoring force is proportional to how far you stretch or compress it.

Okay.

Force depends on displacement.

Now combine that with Newton's second law, F equals ma, force equals mass times acceleration.

Acceleration is the second derivative of position.

Ah, okay.

When you put those two fundamental physical principles together, you naturally get a differential equation describing the spring's motion.

Here.

Specifically, a second order one because of acceleration.

And the type of motion it predicts is a familiar one, isn't it?

That back and forth.

It is.

The solutions for that spring model involve oscillating functions like sine and cosine, which intuitively makes perfect sense for something that bounces up and down, right?

Yeah, it matches what we see.

This demonstrates how differential equations translate fundamental laws of nature into predictable,

observable behaviors.

What's also important to grasp is that differential equations can be of different orders.

First order, second order, depending on the highest derivative involved.

Highest rate of change.

Exactly.

And when we find a function that actually satisfies the equation, that's what we call a solution.

So a solution is essentially the complete historical record or future prediction of that system, right?

Like the whole path it takes.

Precisely.

And often, we're not just looking for any solution, but a particular one that fits a specific initial condition.

An initial condition.

Like a starting point.

Exactly.

Like knowing the precise population at time zero or the exact starting current in an electrical circuit at the moment you flip the switch.

Geometrically, it means picking the unique curve from a whole family of possible solution curves that passes through that specific starting point, t zero y zero.

So it pins down the specific scenario.

Right.

This ability to predict the future state of a system, given its current state, is one of the most powerful aspects of differential equations.

That's brilliant.

But OK, what happens if we can't find a neat, explicit formula for the solution?

I mean, that must happen sometimes.

Is all lost or can we still gain insight?

Not at all.

That's a crucial point.

Often, finding those neat formulas is hard or even impossible.

But when direct analytical solutions are out of reach, we still have powerful graphical and numerical techniques that give us tremendous insight into a system's behavior.

Like direction fields.

That sounds like a visual approach.

Tell me more.

It's remarkably intuitive, actually.

For a given differential equation, say u f x y, it tells you the slope of the solution curve at every point, x y, on a graph.

The slope everywhere.

No.

Well, you can calculate f x y at many points.

So we can draw a grid of tiny line segments, each representing the slope f x y at that specific location.

This collection of segments creates a direction field or slope field.

OK, like little arrows pointing the way.

Exactly.

It acts like a sort of topographical map for solutions, visually guiding us to sketch the general path a solution curve would follow.

You can literally see where solutions would flow, where they'd flatten out, or where they might accelerate just by following the arrows.

And this qualitative understanding is incredibly useful, especially in applied fields like, say, electrical engineering.

Absolutely.

Consider a simple circuit with a resistor R and an inductor.

L.

Kirchhoff's voltage law gives us a first order differential equation for the current i t, L di dt plus R i eq, the voltage.

If we have a constant voltage supply, e, the direction field for this equation, immediately suggests something interesting.

No matter what the initial current is when you start, all solutions will eventually approach a specific limiting current value.

You can see the current settle down to a stable value, an equilibrium, just by looking at the field, even without solving the equation formally.

Wow, OK.

So we can visualize the qualitative behavior even without a precise formula.

But what if we need actual numerical values at specific points in time?

We need numbers sometimes.

That's where Euler's method comes in.

It's a foundational numerical approximation technique.

Euler's method?

OK.

The basic idea is elegantly simple.

You start at your known initial condition, say, x zero, y zero.

The differential equation tells you the slope, y, or direction at that point.

Right, the f x zero, y.

Exactly.

You take a small step, let's call the step size h, in that direction along the tangent line.

Now you're at a new approximate point, y one, y one.

OK, a short straight line jump.

Yep.

At this new location, y one, y one, the equation gives you new slope.

So you adjust your direction based on that slope and take another small step h.

Ah, so you keep correcting your path.

Precisely.

You keep stopping and correcting your direction, like you're navigating step by step using the slope information at each point.

It's like building the solution piece by piece, approximating the curve.

Exactly.

While it doesn't yield an exact formula, by making those step sizes h progressively smaller, you get successively better approximation to the true solution curve.

So smaller steps, better accuracy.

Generally yes.

For instance, if you want to estimate y point three for x plus y starting at y zero one, using a step size of say zero point one, you calculate y zero point one using the slope at zero, then y point two using the slope at point one, and finally y point three using the slope at point two, you might find y zero point three is approximately one point three six two.

OK, so it gives you concrete numbers.

It does.

And this straightforward concept forms the backbone for far more sophisticated computer based numerical methods used in everything from weather forecasting to financial modeling.

It's the starting point for serious computation.

We've explored how to visualize and approximate solutions, which is incredibly powerful.

But for many, I guess, the holy grail is finding that exact formula.

So what kinds of differential equations actually allow us to do that?

When can we solve them perfectly?

You're right.

Finding an explicit formula offers the ultimate precision, that analytical solution.

And there are indeed specific types of differential equations that we can solve directly using integration techniques.

One vital class is separable equations.

Separable, meaning you can separate them.

Exactly.

It means you can, with some algebraic manipulation,

rearrange the equation so that all the terms involving the dependent variable, say y and its differential died, are on one side and all the terms involving the independent variable, say x and dx, are on the other side.

OK, all y is on one side, all x is on the other.

Correct.

Once they're separated, you can simply integrate both sides with respect to the respective variables.

That integration often gives you the relationship between x and y, sometimes explicitly as y equals fx, sometimes implicitly.

It sounds almost like a bit of mathematical magic, where the variables neatly sort themselves out.

It feels that way sometimes.

For instance, d dx is by 2, y2, you can rewrite that as y2 die equals by 2 dx, see?

Separate it.

Ah, yeah.

Then you just integrate both sides.

Integral of y2 tie is y33, integral of y2 dx is by 33, plus a constant of integration c, so y33 equals by 33 plus c, done.

And this applies to real world things.

Definitely.

It's often the case in scenarios like those classic mixing problems.

Ah, mixing problems where you're figuring out how much of a substance like salt is in a tank as liquid flows in and out.

How do differential equations help us there?

Mixing problems are a perfect application of separable equations, often.

Imagine a large tank of water with some salt already dissolved in it.

Then maybe brine flows in at a certain rate in concentration, while the well -mixed solution flows out at some rate.

OK.

If y at t is the amount of salt at time t, then day, the rate of change of salt, is simply the rate salt comes in minus the rate salt goes out.

Rate in minus rate out, makes sense.

You set up that equation based on the flow rates and concentrations.

Often it turns out to be separable.

Solving it then tells you exactly how the amount of salt changes over time.

For example, you might find that after 30 minutes there's roughly 38 .1 kilograms of And maybe in the long run, the amount of salt approaches a stable equilibrium of 150 kilograms.

So it predicts not just the immediate change, but the long -term behavior, too.

Exactly.

It can show a system stabilizing.

But I imagine not all equations are so cooperative and neatly separable, are they?

What if the variables are stubbornly mixed together in a way you just can't pull apart?

That's a great question and very common.

When variables can't be separated, we often check if the equation fits another important

Linear differential equations.

Linear.

Okay, what defines those?

These are first -order equations that can be written in a very specific standard form.

dx plus pxy equals qx.

Here px and qx are functions of only the independent variable x.

All right, y and d at x appear just by themselves, multiplied by functions of x.

Precisely.

The key to solving these linear equations is using something called an integrating factor.

An integrating factor.

That sounds like some kind of secret weapon.

It is, in a way.

It's a special function, let's call it ix, that you would multiply the entire equation by.

This function ix is calculated as e raised to the power of the integral of px dx.

Okay, e to the integral of px.

Why that?

Because the magic is that when you multiply the left side, dx plus pxy, by this specific ix, it transforms perfectly into the derivative of a product.

The derivative of ixy.

Whoa, okay.

It cleans it up.

Exactly.

It uses the product rule in reverse.

So the equation becomes ddx ixy xy xqx, and that is directly integrable.

Ah, so you just integrate both sides now.

Yep.

Integrate both sides with respect to x, and then solve the result for y.

That's a very systematic method.

For example, ddx plus 3 by 2y is 6 by 2.

Here px is 3 by 2.

The integrating factor is e integral of 3 by 2 dx, which is e by 3, got it.

Multiply the whole equation by e by 3, the left side becomes ddx e by 3, the right side is 6 by 2, e by 3.

Integrate, solve for y, and you get yx plus 2 plus c by 3.

So going back to our electric circuit example from earlier, ldi dt plus ri equals et.

That looks like it fits the linear form if we divide by l.

It absolutely does.

gdx plus rli ekl, here pt is rl and qt is edl.

So we can solve it.

Yes.

If the voltage e is constant, the equation is actually both linear and separable, so you have choices.

But if the voltage is variable, say it's oscillating like et equals 6730t, then the equation dit plus ri is linear but not separable.

Aha.

So this is where the integrating factor method becomes essential.

Just nicely.

In that scenario, the integrating factor method is your go -to technique.

It allows us to find an explicit formula for the current I .T.

The solution reveals how the current responds to that oscillating voltage.

It will also oscillate, possibly with a phase shift, and it might have a transient part that dies out over time, showing the complex dynamic interplay between the voltage source and the circuit's properties.

Very cool.

We've covered single populations and simple physical systems like circuits.

But what happens when different species interact within an ecosystem that seems way more complex?

How do differential equations handle that?

That brings us into the fascinating realm of systems of differential equations.

A classic example is modeling predator -prey systems, often described by the Lotka -Volterra equations.

Lotka -Volterra, okay.

These are typically a pair of linked differential equations.

One describes the population of the prey, say rabbits, r, and the other describes the population of the predator, say wolves, w.

They influence each other.

So you have terms representing maybe the rabbits reproducing naturally.

Right, like dr, dt, kr.

But also being eaten by wolves.

Exactly.

Minus a term like arw, where the rate they get eaten depends on how many rabbits and wolves there are.

Okay.

And for the wolves?

Well, maybe they decline naturally if there's no food, like dw, dt, drw.

But their population grows when they eat rabbits.

Precisely.

Plus a term like brw, representing growth proportional to encounters.

The equations directly capture these opposing forces of natural growth, decline, and interaction.

That interaction term rw is key.

It is.

What's particularly interesting here is that for these coupled systems, it's usually impossible to find simple, explicit formulas for rt and wt,

how each population changes over time.

So back to visualization.

Mostly yes.

We analyze them graphically, but in a different way.

We use a phase plane.

Instead of plotting population versus time, we plot the predator population, w, directly against the pre -population, r.

Glee versus r.

What does that show us?

It shows the trajectory of the populations relative to each other.

If time goes on, the point, rt, wt,

traces out a path in this phase plane.

And what kind of patterns do these interspecies interactions create on a phase plane?

Do they just spiral out?

Often no.

In the standard Lotka -Volterra model, the solutions typically form closed cyclical curves.

These cycles revolve around equilibrium points with specific population levels, rw, where both drdt and dwt are zero, meaning the populations could theoretically remain stable if undisturbed.

Equilibrium points, like a balance point.

Exactly.

For example, you might find an equilibrium where, say, 1 ,000 rabbits are just enough to sustain a constant population of 80 wolves indefinitely.

Both rates of change are zero there.

So if we started with, let's say, 1 ,000 rabbits and maybe only 40 wolves, what would we observe happening to their numbers on this phase plane?

You'd see the point, rw, start to trace out one of those closed curves.

A beautiful, predictable cycle.

How does it cycle?

Well, starting with 1 ,000 rabbits and only 40 wolves.

Initially, with fewer wolves, the rabbit population would increase.

The point moves to the right.

More rabbits.

More rabbits mean more food for the wolves, so the wolf population, w, would then begin to increase.

The point moves up.

Okay, right.

And then up.

As the wolves become abundant, they consume more rabbits, causing the rabbit population r to decline.

The point moves left.

Left.

Ah, fewer rabbits.

Fewer rabbits, in turn, lead to less food for the wolves, causing their population w to decline.

The point moves down.

Down.

And that brings us back.

Back towards where we started, completing the cycle.

The populations oscillate, with the prey population typically peaking a bit before the predator population.

They're always chasing each other in this continuous, dynamic loop represented by that closed curve in the phase plane.

And this isn't just a theoretical construct, is it?

This cycling pattern, has it been observed in the wild?

Oh, absolutely.

It's not just theoretical.

Historical records.

Famously, the Hudson's Bay Company's fur trade data from Canada over many decades.

Ah, yes.

The links in the hair.

Exactly.

That data shows precisely these kinds of coupled oscillations between snowshoe hair, prey, and Canada links, predator populations, often with roughly a 10 -year period.

It's a really powerful, real -world validation of these mathematical models.

Amazing.

And it underscores their ability to explain complex ecological dynamics.

Of course, these basic models can also be expanded and refined to include other factors like environmental changes,

resource limitations for the prey, or even harvesting by humans.

Wow, we've really covered a lot of ground today.

From modeling the microscopic world of bacteria to springs, circuits, and all the way up to the majestic dance between predators and prey in an ecosystem.

It truly highlights the incredible versatility and, frankly, the predictive power of differential equations.

They are, in essence, the mathematical language of change.

They allow us to quantify, predict, and understand phenomena across biology, physics, engineering, economics, and so much more.

Hopefully, you've now gained a solid conceptual understanding of how these equations are formulated, how we can visualize their solutions, approximate them numerically, and sometimes solve them exactly, and maybe most importantly, why they matter.

So what does this all mean for you listening right now?

Well, next time you read a new story about a population trend, a medical breakthrough involving drug concentration over time, or even a new piece of technology, consider for a moment how a differential equation might be working behind the scenes at its very core.

It empowers you to ask deeper questions.

That's right.

Like,

how would changing the initial conditions of a system, that starting point, dramatically alter its long -term behavior?

Or think about this.

If a rocket is fired straight up from Earth, we can use differential equations involving gravity and air resistance to figure out how long it would take to reach its maximum height.

But would it take the same amount of time to fall back down?

Why or why not?

Interesting.

Air resistance would likely make the fall longer.

Exactly.

These are the kinds of thought -provoking questions that differential equations not only allow

Thank you for joining us on this deep dive from the Last Minute Lecture team.