Chapter 11: Sequences, Series, and Power Series

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.

You know that feeling when you're just buried under information?

Stacks of articles, notes, research papers, and you just need the key takeaways like now.

You want to be properly informed but without drowning in details.

Yeah, absolutely.

And that's basically what we do here on the deep dive.

We try to sift through it all, find those really crucial bits of knowledge, and make even the complicated stuff more engaging so you get those proper aha moments.

Today, we're taking a deep dive into something pretty core to calculus.

Chapter 11 of calculus, early transcendentals by Stewart, Clegg, and Watson.

It's all about sequences, series, and power series.

Sounds maybe a bit heavy.

It can sound that way, but honestly, the concepts are incredibly elegant and surprisingly useful.

Our mission today is really to break these ideas down step by step.

Look at how they work, why they're important, and crucially, where they pop up in the real world.

Right, like how this math actually connects to things like physics or biology.

Okay, so let's get started.

The very first building block, sequences.

What exactly is sequence?

Well, in simple terms, it's just an ordered list of numbers.

Think of it like a lineup, a first number, a second, a third.

You get the idea.

A1, A2, A3, and so on.

Okay, an ordered list.

You can also think of it like a function, actually.

A function where your input is just the counting numbers 1, 2, 3,

and the output is the number at that specific position in the list.

So you might have a formula like a 1 over 2 to the n that gives you 12, 14, 18.

Exactly, that's one way.

Or sometimes a term is defined by the terms that came before it.

That's called a recursive definition.

Oh, okay.

And to get a handle on what an infinite list really means, maybe we can start with something classic, like Zeno's paradox.

Perfect starting point.

Yeah, Zeno's paradox.

Imagine someone trying to walk to a wall.

First, they walk half the distance, then they walk half of the remaining distance.

And then half of what's left after that and so on forever.

Right.

So the distances covered in each step form this sequence.

12, 14, 18, 116.

It's an infinite list generated by this simple scenario.

And this immediately brings up that weird question.

They're always moving, covering smaller and smaller bits of ground, but do they ever actually get to the wall?

It feels like they shouldn't, maybe.

That question gets right to the heart of why sequences and later series are so interesting.

They let us analyze these infinite processes.

And sequences aren't just formulas.

Think about, say, the world population recorded each year.

That's a sequence.

Or the digits of pi, 3 .14159.

That's an infinite sequence too, right?

No simple formula there.

Exactly.

And then there's the Fibonacci sequence, probably the most famous recursive one, starts one, one, that each next term is the sum of the two before it.

One, one, two, three, five, eight.

It came from like rabbits breeding or something.

Yeah, a 13th century problem about rabbit populations.

But the amazing thing is how often it appears in nature.

Pine cone spirals, arrangement of leaves, that kind of stuff.

So when we have these sequences, what's the big question we ask about them?

A really crucial question is,

do the numbers in the list eventually kind of settle down?

Do they get closer and closer to one specific value?

That's the whole idea of a limit.

Okay, so convergence.

Right.

We say a sequence converges to a limit.

Let's call it L.

If its terms get arbitrarily close to L as you go further and further out in the sequence, imagine plotting them on a number line.

They'd bunch up around L.

And if they don't settle down, if they just bounce around or shoot off to infinity.

Then we say the sequence diverges.

It doesn't approach one single finite number.

And there are useful rules too.

Like if you have two convergent sequences, you can add them, subtract them, multiply them.

And the resulting sequence also converges in a predictable way.

Okay, that makes sense.

So sequences are these ordered lists.

Now what happens if we decide to

add up all the numbers in an infinite list?

Now you're making the jump from sequences to series.

An infinite series is what you get when you try to sum the terms of an infinite sequence.

So back to Zeno's runner.

Exactly.

The question, does the runner reach the wall, is mathematically the same as asking.

Does the infinite sum 12 plus 14 plus 18 plus 116 plus phones actually add up to a specific finite number?

And intuitively we know the runner does reach the wall, eventually covering the whole distance, which is one.

So it seems possible for infinitely many numbers to add up to something finite.

Which still feels a bit weird.

You can't actually sit there and add them all up, right?

No, you definitely can't perform infinite additions.

So the way mathematicians handle this is with the concept of partial sums.

Think of it like this.

You add the first term, that's the first partial sum, then add the first two terms, second partial sum, then the first three, and so on.

Okay, so you get a sequence of sums.

Precisely.

You get a sequence of these partial sums, S1, S2, S3.

And the question becomes, does this new sequence, the sequence of partial sums, converge to a limit?

Ah, okay.

So if the sequence of partial sums converges to some number, let's say S.

Then we say the original infinite series converges, and its sum is that number S.

If the sequence of partial sums diverges, then the series diverges.

It doesn't add up to a finite value.

Are there examples where this summing process does something, like, unexpected?

There are.

There's a neat type called the telescoping sum.

Imagine the series where each term is one divided by n times n plus one.

So one twelve plus one twenty three plus one thirty four plus sum.

Okay.

If you write out the partial sums using a little algebraic trick called partial fractions, you find that loads of intermediate terms cancel each other out.

It collapses, like an old spyglass or telescope.

And what does it collapse to?

Amazingly, the sum of that infinite series is just one.

Very clean.

Wow.

Okay.

But probably the most important type of series we need to know is the geometric series, right?

Absolutely fundamental.

A geometric series has the form

A plus R plus R2 plus R3 plus egg.

You start with A, and then you multiply by the same number R, the common ratio, over and over again.

And Zeno's paradox sum, twelve plus fourteen plus eighteen, that's one of these, isn't it?

Perfect example.

Here, A is twelve, and the common ratio, R, is also twelve.

And here's the really powerful result.

If the absolute value of R is strictly less than one.

So R is between minus one and one, but not including minus one or one.

Right.

If R is one, the geometric series converges.

And even better, it converges to a very simple formula.

A one R.

That's incredibly useful.

So for Zeno A12 R12, the sum is twelve one twelve R12, just like we expected.

Exactly.

But if R is one or greater, the series diverges.

It either adds up terms that don't get smaller or they get bigger.

And this geometric series formula pops up in practical places?

Oh yeah.

Calculating drug concentration in the body after repeated doses, figuring out the multiplier effect in economics, even converting repeating decimals like say 2 .3177777 into a fraction.

It's a workhorse.

Okay, so geometric series are nice and predictable if R one.

What about series that are less well behaved?

Well, let's talk about the harmonic series.

One plus twelve plus thirteen plus fourteen plus fifteen plus our and the terms one N definitely gets smaller and smaller.

They approach zero.

So you'd kind of think it should converge, right?

Adding smaller and smaller bits.

That's the intuition trap.

But no, the harmonic series actually diverges.

It grows infinitely large, just very, very slowly.

Nicole Resmay proved this way back in the 14th century by grouping terms cleverly to show the sun keeps exceeding any finite number.

That's really counterintuitive.

It's a crucial warning sign, isn't it?

Just because the terms go to zero doesn't guarantee convergence.

Precisely.

And that leads us directly to the very first test you should always apply to a series.

The test for divergence.

Okay, what's that?

It's based on the flip side of what we just learned.

Theorem six in the book states, If a series converges, then its terms must approach zero as N goes to infinity.

Right.

If you're adding up infinitely many things and getting a finite sum, the things you're adding better be getting vanishingly small.

Exactly.

So the test for divergence uses the contrapositive.

If the limit of the terms as N approaches infinity is not zero, or if the limit doesn't even exist, Then the series cannot converge.

It must diverge.

Correct.

It's a quick knockout test.

If you look at a series and its terms don't even go to zero, you're done.

It diverges.

No need for more complicated tests.

But again, the big warning.

If the terms do go to zero, this test tells you absolutely nothing.

The series might converge or it might diverge like the harmonic series.

That's the crucial takeaway.

The test for divergence can only tell you if a series diverges.

It can never confirm convergence.

So if the terms do go to zero and it's not a geometric or telescoping series where we have formula,

how do we figure out convergence then?

Finding formulas for partial sums sounds hard for most series.

It's usually impossible for most series.

That's why we need a whole arsenal of other convergence tests.

Think of them as different diagnostic tools.

Okay.

Like what?

One really elegant one is the integral test.

It connects the convergence of a series with positive decreasing terms to the convergence of an improper integral.

How does that work?

Imagine the terms of the series as the heights of rectangles of width one.

The sum of the series is the total area of these rectangles.

Now find a continuous positive decreasing function f x that matches the series terms at the integers.

So f n equals n.

Okay.

Like a curve that goes through the top corners of the rectangles.

Exactly.

The integral test says that the infinite series is a and fun a converges if and only if the improper integral of x from one to infinity converges, meaning the area under the curve is finite.

So infinite sum behaves like infinite area under the related curve.

That's neat.

Does it give you the sum?

Ah, good question.

No, usually not.

The value of the integral is generally not the same as the sum of the series.

For example, the series one and two converges.

We know this from the integral test comparing to one by two d x and its sum is famously 26.

But the integral from one to infinity is just one.

They converge together, but often to different values.

Okay.

So it's just for telling convergence and this helps with those p series.

Yes.

The integral test is perfect for analyzing the p series, which is a one and p.

It tells us this series converges if p one and diverges if p one.

This gives us a whole family of known convergent or divergent series to compare against.

Comparing.

That sounds like another type of test.

It is.

We have the comparison tests.

The direct comparison test is pretty intuitive.

If you have a series with positive terms and you can show its terms are smaller than the convergent series, like a p series with p one, then your series must also converge.

Makes sense.

Smaller than finite is finite.

And conversely, if your terms are larger than the terms of a known divergent series, like the harmonic series p one, then your series must also diverge.

Bigger than infinite is infinite.

What if direct comparison is tricky?

Then you use the limit comparison test.

You take the limit of the ratio of the terms of your unknown series and a known series.

If that limit is a finite positive number, then both series do the same thing.

Either both converge or both diverge.

It's often easier than setting up inequalities.

Okay.

So those handle series with positive terms.

What about series where the signs flip back and forth?

Those are alternating series, like 1, 12 plus 13, 14 plus nair.

That's the alternating harmonic series.

For these, we have the alternating series test.

What does that say?

An alternating series converges if two simple conditions are met.

First, the absolute value of the terms is decreasing.

Ignoring the sign, the numbers get smaller.

Second, the terms themselves approach zero.

So the alternating harmonic series, 1, 12 plus 13, its terms decrease an absolute value, 1, 12, 13, and they go to zero.

So it converges.

It does.

Unlike the regular harmonic series, the alternating signs give it just enough cancellation to converge.

Visually, the partial sums oscillate back and forth, but the swings get smaller and smaller, homing in on a limit.

And there's something cool about estimating the sum for these, right?

Yeah, that's the alternating series estimation theorem.

It's incredibly useful.

It says that if you approximate the sum of a convergent alternating series by using just the first n terms, a partial sum, the error you make the difference between your approximation and the true infinite sum is less than the absolute value of the first term you left out, the n plus 1th term.

Wow.

So you know exactly how good your approximation is just by looking at the next term.

That's handy.

It really is.

Now, this brings up a deeper point about convergence.

Yeah.

Absolute versus conditional convergence.

Okay, what's the difference?

A series is absolutely convergent if the series you get by taking the absolute value of every term also converges.

So n it converges.

So it converges even without the help of alternating signs.

Exactly.

Absolute convergence is stronger, more robust.

A series is conditionally convergent if the series itself converges, but the series of its absolute values diverges.

Like the alternating harmonic series, it converges, but the harmonic series, its absolute values diverges.

Precisely.

And here's the mind -blowing part.

Ryman proved that if a series is conditionally convergent, you can actually rearrange the order of its terms to make it sum up to any real number you want, or even make it diverge.

Seriously, rearranging the terms changes the sum?

That doesn't happen with finite sums?

Nope.

It's a purely infinite phenomenon.

Absolute convergence behaves like finite sums.

Rearranging doesn't change the sum.

Conditional convergence is much more delicate.

Wild.

Okay, any other major tests in the toolkit?

Two more heavy hitters.

The ratio test and the root test.

The ratio test is often the go -to for series involving factorials like n or nth powers of constants like 2n.

How does it work?

You look at the limit of the absolute value of the ratio of consecutive terms.

An plus 1n.

Let's call this limit l.

If l1, the series converges absolutely.

If l1 or infinity, the series diverges.

And if l equals 1?

Then the test is inconclusive.

It tells you nothing.

You have to try a different test.

This often happens with p -series, for example.

The ratio test fails there.

Okay.

And the root test?

The root test is similar, but you look at the limit of the nth root of the absolute value of the nth term.

Limit of an o1.

Call that limit elegant.

The conclusions are the same.

l1 means absolute convergence.

l1 means divergence.

l1 is inconclusive.

When would you use the root test?

It's particularly handy when the entire nth term is raised to the nth power, like in at 2n place 1, 3n.

Taking the nth root simplifies things nicely.

Got it.

So a whole suite of tests to tackle different kinds of series.

Now,

where does this all lead?

What's the ultimate application of all this infinite series stuff?

It leads us to perhaps the most powerful tool in the chapter.

Power series.

We're shifting from series of constant numbers to series that involve a variable, usually x.

So instead of just numbers, we have terms like c0 plus c1x plus c2 by 2 plus c3 by 3 plus f star.

Exactly.

A power series is essentially an infinitely long polynomial.

And the amazing thing is a power series defines a function.

The function fx equals the sum of the series.

But presumably it only works for certain values of x, right?

Like the geometric series.

Precisely.

The set of x values for which the power series converges is called its interval of convergence.

For every power series centered at a, there's a radius of convergence, r.

The series converges absolutely for x ar and diverges for x ar.

So it converges inside an interval, ar a plus r.

What about at the endpoints x ar and x was a plus r?

You have to check those separately.

The ratio and root tests usually give you r, but they are inconclusive at the endpoints.

You plug those specific x values back into the series and use one of the other tests, like p series, alternating series test, et cetera, to see if it converges or diverges right there.

Okay.

So power series define functions on these intervals.

What's so special about them?

Here's the magic.

And it's truly remarkable.

Within their interval of convergence, excluding the endpoints usually, you can differentiate and integrate power series term by term, just as if they were regular polynomials.

Wait, you can just differentiate or integrate the infinite sum term by term.

Yep.

And the resulting new power series will have the same radius of convergence as the original one.

That seems incredibly powerful.

Why is that so useful?

Think about functions whose integrals we couldn't find before, like e by two, which is crucial in statistics.

We can't write down a simple function for its integral, but we can write down a power series for e by two.

And then we can integrate that power series term by term to get a power series for its integral.

Isaac Newton grasped this power early on.

So you can find series for things you couldn't integrate otherwise.

Exactly.

Or you can find new series from old ones.

Differentiate the series for one, one X, which is XN, and you get the series for one, one by two, integrate the series for one, one plus by two, and you get the series for arc tan X.

And isn't the arc tan series related to calculating pi?

It is.

Plugging X one into the arc tan X gives the famous Leibniz formula for pi four, 113 plus 15, 17 plus 17.

Slow convergence, but beautiful.

Okay, this leads us to the grand finale, right?

Taylor and McLaurin series.

Are these just special kinds of power series?

They are the ultimate power series representation of functions.

The Taylor series gives us a way to represent almost any sufficiently smooth function as a power series centered around a specific point A.

How does it do that?

Where do the coefficients come from?

This is the genius part.

The coefficients are determined by the function's own derivatives evaluated at the center point A.

Specifically, the coefficient of XAN is the nth derivative of F at A divided by N factorial, FNAN.

Wow.

So the function's local behavior at A, its value, its slope, its curvature, et cetera, dictates the entire infinite series representation.

Exactly.

And a McLaurin series is simply a Taylor series centered at Egalizano.

This often gives the simplest form.

And can we find these for common functions?

Oh, yes.

The McLaurin series for functions like X, SYNX, and COSEX are incredibly elegant and fundamental.

For instance, X, X1 plus X plus by 22 plus by 33 plus S, which is pexing, and then SYNX and COSEX have similar beautiful patterns involving only odd or even powers respectively.

So if you find a power series for a function, say by manipulating a geometric series or integrating, is that automatically its Taylor series?

Yes.

A function can only have one power series representation centered at a given point.

So however you find it, that is the Taylor or McLaurin series.

This means you can often find Taylor series by clever substitution, multiplication, differentiation, or integration of known series without calculating all those derivatives directly.

That saves a lot of work.

You mentioned the binomial series, too.

Right.

The binomial series is the Taylor series for 1 plus XK, where K can be any real number, not just a positive integer.

It generalizes the binomial theorem and converges for the 1.

Super useful for approximations involving roots or fractional powers.

Okay, so we have these amazing infinite series representations, but in the real world, computers can't compute infinite things.

How are these actually used?

Great question.

In practice, we often use the finite partial sums of Taylor series.

These are called Taylor polynomials.

A Taylor polynomial Tn on X is just the first few terms of the Taylor series, up to the term with Xan.

So you're using a polynomial to approximate the original function.

Precisely.

And polynomials are wonderful because they're incredibly easy for computers and calculators to evaluate.

Just addition, subtraction, multiplication.

When you press the sin button or the X button on your calculator, it's likely computing a highly accurate Taylor polynomial approximation very quickly.

And we know how good these approximations are.

Yes, that's where Taylor's inequality comes in.

It provides a rigorous way to put an upper bound on the error or remainder when you approximate a function with its Taylor polynomial.

This allows engineers and scientists to guarantee a certain level of accuracy for their calculations.

This really sounds like where the math meets the real world, especially in physics and engineering.

Absolutely.

Taylor polynomials are indispensable there.

Let's take Einstein's special relativity.

The formula for kinetic energy gets complicated at speeds near the speed of light.

But if you take the relativistic kinetic energy formula and find its Taylor polynomial approximation for low velocities, V much smaller than C,

the first main term you get is exactly the classical Newtonian kinetic energy, 12 millivit.

Whoa.

So the Taylor series shows how Newton's familiar physics is actually embedded within Einstein's theory as an approximation for everyday speeds.

Exactly.

It provides that beautiful bridge showing the more general theory reduces to the older one in the appropriate limit.

It's not that Newton was wrong, just that his theory was an excellent approximation under normal conditions.

That's incredibly cool.

What about other areas like optics?

In optics, designing lenses involves complex equations with trigonometric functions like sine and cosine.

Physicists often use the first degree Taylor approximations like sin theta, cos theta, one for small angles, to simplify these equations dramatically.

This leads to Gaussian optics or paraxial optics, the basic theory used for simple lens design.

So simplifying the trig functions with their Taylor polynomials makes the physics manageable.

Yes.

And if you need more accuracy, like for high quality camera lenses, you use higher order Taylor polynomial approximations, like including the cubic term for sine, which leads to more refined theories like third order optics.

Are there other quick examples where this simplification is key?

Loads.

Analyzing electric fields far from a dipole, understanding water wave velocity in shallow water, correcting for earth's curvature and highway design, figuring out the period of a pendulum when it swings wide.

In all these cases, replacing a complex function with the first few terms of its Taylor series gives crucial insights and workable models.

You mentioned black body radiation too.

Right.

A huge moment in physics history.

Classical physics, the Rayleigh -Jeans law, failed badly for black body radiation at short wavelengths, the ultraviolet catastrophe.

Max Planck came up with a new formula, Planck's law, involving quantum ideas.

And Taylor series connect them.

Yes.

If you take Planck's law and find its Taylor approximation for long wavelengths, it perfectly reduces back to the classical Rayleigh -Jeans law.

Again, showing how the new correct theory contains the old one as a limiting case.

So the core idea running through all these applications is using Taylor polynomials to simplify complexity and gain understanding.

Replacing messy functions with manageable polynomials to see the underlying behavior.

That's it.

Exactly.

It's one of the most powerful practical techniques derived from calculus.

What a journey.

We've gone from simple ordered lists and sequences to the tricky idea of summing infinitely many numbers in series, developed a whole toolkit of convergence tests, then built functions out of infinite polynomials with power series, and finally saw how Taylor series and polynomials let us approximate complex reality and unlock secrets in physics and engineering.

It really covers a lot of ground, building idea upon idea.

We genuinely hope this deep dive has made these fundamental calculus concepts feel clearer, maybe more intuitive, and definitely more engaging.

Hopefully you had a few of those aha moments we talked about.

Yeah, we hope so too.

So here's a thought to leave you with.

Considering how powerful these polynomial approximations are, what other really complex phenomena out there, maybe in biology, economics, or even deeper in physics, might just be waiting for us to uncover their simpler, underlying polynomial nature using these very tools.

Something to think about.

Thanks so much for joining us on this deep dive into calculus chapter 11.

Until next time, keep exploring.

From everyone here at the Last Minute Lecture Team, we really appreciate you listening.