The one core idea: if an equation ties a curve to its own slope and its own bending, we can guess the curve is an endless polynomial a 0 + a 1 x + a 2 x 2 + ⋯ and let the equation compute the unknown numbers one at a time. This page builds — from nothing — every symbol you must already own before that guess makes any sense: what a power, a sum-sign, a slope, a bend, and a "nice function" actually are .
This is the toolbox page for Power series solutions at an ordinary point . We assume zero prior notation. Read top to bottom; each block earns the symbols the next block spends.
A variable is just a number we're allowed to slide, like a bead on a wire. We call it x . A function y is a rule: give it an x , it hands back a height y . We write y ( x ) — "the height at position x ."
Picture: a horizontal number-line for x , a vertical number-line for the height y , and a curve floating above them. Pick any x , walk straight up to the curve, and the height you reach is y ( x ) .
Why the topic needs it: the whole game is finding an unknown curve y ( x ) . Everything else is machinery for pinning down that curve.
x n means "multiply x by itself n times." So x 2 = x ⋅ x , x 3 = x ⋅ x ⋅ x . Two edge cases you must never forget:
x 1 = x (one copy, itself),
x 0 = 1 (zero copies = the "do nothing" number, 1 ).
Picture the shapes: x 0 = 1 is a flat horizontal line. x 1 is a straight ramp. x 2 is a U-shaped bowl (a parabola). Higher powers get steeper and more curved. Each power is a different building-block shape .
Why the topic needs it: a power series is a recipe that mixes these shapes . To trust the recipe you must first see each ingredient shape on its own.
"x 0 = 0 , surely?"
Why it feels right: zero exponent, so zero?
The fix: x 0 = 1 . Think of multiplying by "no copies of x " as leaving 1 untouched. This matters because the very first term of every series, a 0 x 0 , is just the constant a 0 .
A coefficient is the amount of each building-block shape we pour into the mix. a n (read "a-sub-n") is the amount of the x n shape. The little n underneath is a subscript — a name-tag, not a multiplication.
Picture: a row of tuning dials labelled a 0 , a 1 , a 2 , … . Each dial controls how much of one shape (1 , x , x 2 , …) is added. Turn the dials and you sculpt any curve you like.
Why the topic needs it: the ODE method never guesses the shape of the answer — the shapes (x n ) are fixed. It only solves for the dial settings a n . Those numbers ARE the solution.
n = 0 ∑ ∞ a n x n is shorthand for an endless addition:
∑ n = 0 ∞ a n x n = a 0 x 0 + a 1 x 1 + a 2 x 2 + a 3 x 3 + ⋯
Read it: "start the counter n at 0 , write the term a n x n , bump n up by one, keep going forever (∞ = never stops)." The n is a dummy index — a temporary counter; renaming it n → k changes nothing.
Picture: a conveyor belt. Slot n = 0 drops a 0 , slot n = 1 drops a 1 x , slot n = 2 drops a 2 x 2 … The ∑ is just "sum whatever the belt delivers."
Why the topic needs it: writing a 0 + a 1 x + a 2 x 2 + ⋯ by hand forever is impossible. The ∑ lets us manipulate the whole infinite list at once — differentiate it, shift it, add two of them.
"Changing the lower limit is optional."
Why it feels right: it's just a counter.
The fix: the lower number tells you where the first real term sits. Later, when we shift indices, the lower limit shifts too — get it wrong and you lose or duplicate a term. This is THE classic re-indexing bug in the parent note.
A power series about x 0 = 0 is exactly n = 0 ∑ ∞ a n x n : an "infinite polynomial." A function is analytic at a point if, near that point, it equals some convergent power series — i.e. the infinite sum actually adds up to a finite height, and that height is the function.
Picture: take the U-shaped x 2 , the ramp x , the flat 1 , and infinitely more, each scaled by its dial a n , and stack them. If they settle down to a smooth finished curve (instead of blowing up), the function is analytic. Analytic = "reachable by mixing power shapes."
Why the topic needs it: the parent note defines an ordinary point as one where the coefficients P ( x ) , Q ( x ) are analytic. That word is the gatekeeper of the whole method. See Taylor series and analyticity for the deeper story of which dials to set.
"Any function is analytic — just take a Taylor series."
Why it feels right: smooth-looking curves seem harmless.
The fix: a function like 1 − x 2 1 blows up at x = ± 1 , so its series only "works" (converges) close to 0 . Analytic is a local badge tied to a point — that's exactly why the topic tracks the nearest trouble spot.
The derivative y ′ ( x ) (read "y-prime") is the slope of the curve y at position x : how steeply it climbs or falls right there. Uphill ⇒ y ′ > 0 ; downhill ⇒ y ′ < 0 ; a flat spot ⇒ y ′ = 0 .
Picture: lay a tiny straight ruler flush against the curve at one point — the tilt of that ruler is y ′ . Slide along the curve and the ruler's tilt changes; that changing tilt is the function y ′ ( x ) .
Why the topic needs it: an ODE is a sentence about slopes . To plug our series guess into that sentence, we must know the slope of the guess — and the rule above turns "slope of a series" into "another series."
y ′′ (read "y-double-prime") is the slope of the slope — how the steepness itself is changing. It measures curvature : how sharply the curve bends. Bowl-up ⇒ y ′′ > 0 ; dome-down ⇒ y ′′ < 0 ; straight ⇒ y ′′ = 0 .
Picture: the U-bowl x 2 curves upward everywhere — its y ′′ is positive throughout. A dome curves downward — negative y ′′ . A straight ramp doesn't bend — y ′′ = 0 .
Why the topic needs it: the equations we solve are second-order — they involve y ′′ . The parent's series for y ′′ , ∑ n ≥ 2 n ( n − 1 ) a n x n − 2 , is nothing but this drop-rule applied to every term at once.
An ODE (ordinary differential equation) is one equation that mixes a curve y , its slope y ′ , and its bend y ′′ , and demands they balance to zero. P ( x ) and Q ( x ) are known coefficient functions that weight the slope and the curve. "Linear " means y , y ′ , y ′′ appear only to the first power — no y 2 , no y y ′ .
Picture: at every point the curve must satisfy a local budget — "(its bend) plus P × (its slope) plus Q × (its height) = 0 ." The curve has to bend in exactly the way that keeps this budget balanced everywhere.
Why the topic needs it: this is the object we solve. "Standard form " means the y ′′ term has coefficient exactly 1 — you get there by dividing through by the leading coefficient, which is precisely how the hidden singular points (where you'd divide by zero) reveal themselves. Guarantees that a solution even exists live in Existence and uniqueness for linear ODEs .
A recurrence relation is a rule that computes each new coefficient from earlier ones, e.g. a k + 2 = − ( k + 2 ) ( k + 1 ) a k . Feed it a starting number, it hands back the next, forever.
Picture: dominoes. Knock over a 0 and the rule topples a 2 , then a 4 , then a 6 … A separate line of dominoes runs a 1 → a 3 → a 5 → ⋯ . The two lines never touch — which is exactly why a 0 and a 1 are free and give the two independent solutions.
Why the topic needs it: substituting the series turns the differential equation into this algebraic domino rule. That trade — calculus for arithmetic — is the entire payoff. The general anatomy of such rules lives in Recurrence relations .
variable x and function y of x
summation sign builds a power series
analytic function equals its series
derivative y prime is slope
second derivative y double prime is bending
ordinary point needs P and Q analytic
the ODE y double prime plus P y prime plus Q y equals zero
substitute then recurrence relation
Power series solution y1 and y2
This map feeds forward into the parent method and onward to Legendre's equation and Legendre polynomials , Hermite and Airy equations , and (when a point is not ordinary) Singular points and Frobenius method .
Cover the right side and answer each aloud — if any stumps you, re-read that section before the parent note.
What does x 0 equal, and what shape is it? x 0 = 1 ; a flat horizontal line.
What does the subscript in a n mean? It's a name-tag: a n is the amount of the x n shape, not a multiplication.
Expand ∑ n = 0 2 a n x n in full. a 0 + a 1 x + a 2 x 2 .
Why does the y ′′ series start at n = 2 ? Because d x 2 d 2 x 0 = 0 and d x 2 d 2 x 1 = 0 — constants and ramps have no bending, so their terms vanish.
Differentiate x n once, then twice. n x n − 1 ; then n ( n − 1 ) x n − 2 .
In plain words, what does y ′ measure? What does y ′′ measure? y ′ is the slope (steepness); y ′′ is the curvature (how the slope bends).
What does "analytic at a point" mean? Near that point the function equals a convergent power series.
What turns the differential equation into arithmetic? Substituting the series produces a recurrence relation — an algebraic rule linking the coefficients.
Why are a 0 and a 1 free? The recurrence links a k + 2 to a k only, so the two starting dominoes (a 0 even-line, a 1 odd-line) are never fixed by it.