What the Laplace Transform Does

A bridge to an easier world

By now you can solve a constant-coefficient equation a y'' + b y' + c y = 0 by guessing y = e^(rt) and reading off the roots of its characteristic equation. That works beautifully when the right-hand side is zero or a tidy forcing term. But what about a circuit that gets switched on at t = 3, or a beam struck by a sudden hammer blow? The Laplace transform is the tool built for exactly those rough, switched, impulsive inputs — and, as a bonus, it folds the initial conditions right into the algebra instead of leaving them for a final clean-up step.

Here is the central image. Solving an ODE in the world of t — the time domain — means doing calculus: differentiating, integrating, fighting with derivatives. The Laplace transform carries your problem across a bridge into a new world, the s-domain, where each function of t becomes a function of a new variable s. The magic is that on the far side, the operation of differentiation turns into plain multiplication. Calculus on this bank becomes algebra on that bank. You solve the easy algebra problem over there, then carry the answer back across the bridge.

The integral that does the carrying

So what is this bridge, concretely? The Laplace transform of a function f(t) is an integral that eats f and produces a new function, written F(s) or L{f}: F(s) = integral from 0 to infinity of e^(-s t) * f(t) dt. Read it slowly. You multiply your function by a decaying weight e^(-s t), then add up the result over all time from 0 onward. The output no longer depends on t — that variable has been integrated away — it depends only on s. One function of t in; one function of s out.

Definition:   L{f}(s) = F(s) = integral_0^infinity  e^(-s t) f(t) dt

f(t) = 1      ->  F(s) = 1/s            (s > 0)
f(t) = e^(a t) ->  F(s) = 1/(s - a)      (s > a)
f(t) = t      ->  F(s) = 1/s^2          (s > 0)

Linearity:    L{ a f + b g } = a L{f} + b L{g}

Try the simplest case by hand to see it is no mystery. Take f(t) = 1. The integral is integral from 0 to infinity of e^(-s t) dt, which is the area under a decaying exponential — equal to 1/s, provided s > 0 so the tail actually decays. So L{1} = 1/s. The same one-line calculation with f(t) = e^(a t) gives 1/(s - a). You are not memorising spells; each entry in the table is just this integral, computed once. Next guide collects the entries you will reuse so you never recompute them.

Two rules that make it usable

An integral defined one function at a time would be tedious. Two properties rescue it. The first is linearity: the transform of a sum is the sum of the transforms, and constants pull straight out — L{a f + b g} = a L{f} + b L{g}. This is immediate, because the defining integral is itself linear. It means you can transform a complicated combination like 3 - 2 e^(t) + 5 t piece by piece, look each piece up, and add. The whole method inherits the same linearity that made linear equations tractable in the first place.

The second property is the one that earns the transform its place in this rung: the transform of a derivative is L{y'} = s * Y(s) - y(0), where Y is the transform of y. Look at what happened. Differentiation in t became multiplication by s in the s-domain — exactly the calculus-into-algebra promise — and the initial value y(0) dropped out of the formula as a bonus term. Apply it twice and L{y''} = s^2 * Y - s * y(0) - y'(0), now carrying both y(0) and y'(0). The next guide is devoted entirely to this rule and the way it builds initial conditions in.

There and back: the round trip

A one-way bridge would be useless. Once you have wrestled the algebra in the s-domain and hold an expression Y(s), you need to come home — to recover the y(t) it represents. That return trip is the inverse Laplace transform, written L^(-1). In principle it is a contour integral in the complex plane, but you will almost never compute it that way. In practice you run the transform table backwards: if you can rewrite Y(s) as a sum of pieces you recognise from the table, the inverse of each piece is read off directly.

The trouble is that the Y(s) you get from an ODE usually arrives as a single messy fraction — a ratio of polynomials in s — that matches no table entry as written. The cure is partial fractions, the algebraic tool that splits one ugly fraction into a sum of simple ones, each of which the table knows. For example, 1/((s-1)(s-2)) is not in the table, but it equals -1/(s-1) + 1/(s-2), and each of those inverts to an exponential. Guide four is built around this exact splitting step, because it is where most of the real arithmetic in the method lives.

The whole method in one breath

Now stand back and see the shape of the complete plan — the master algorithm that the rest of this rung fills in. It is a there-and-back-again journey with the hard part replaced by algebra in the middle. What makes it special compared with the characteristic-equation route you already know is the second word in the next callout: the initial conditions are not bolted on at the end, they are transferred into the algebra from the very first step.

1. Transform both sides of the ODE, using L{y'} = sY - y(0) so the initial conditions enter immediately.
2. Solve the resulting algebraic equation for Y(s) — no calculus, just rearrange.
3. Break Y(s) into table-sized pieces with partial fractions.
4. Invert each piece with the table to recover y(t), the answer in the time domain.

That is the entire arc of the rung in four moves: transform, solve, split, invert. Notice you never solve a differential equation in the usual sense at all — you transform it into algebra, do the algebra, and transform back. The remaining guides simply equip you for each move: a working table to make step one and step four fast, the derivative rule that powers step one, the partial-fraction craft for step three, and finally a full initial value problem carried end to end. Keep the bridge picture in mind and none of it will feel like a bag of tricks.