JOVANA
Library Glossary Getting Started Three Levels Fields How it works Mission
Join the mission
All guides

Transforming Derivatives: Initial Conditions Built In

One short rule, L{y'} = sY - y(0), is the engine of the whole method: it turns differentiation into multiplication and pulls the initial conditions straight into the algebra. See where it comes from, why y(0) falls out, and how it converts an entire ODE into an equation you can solve by hand.

The rule that does all the work

Guide one promised it and guide two stocked the table around it; now we meet the single fact that makes the Laplace transform worth learning. It is the transform of a derivative: if Y(s) is the transform of y(t), then L{y'} = s * Y(s) - y(0). Read it as a trade. On the left is a derivative — a calculus operation. On the right there is no derivative at all: just a multiplication by s and the subtraction of one number, the starting value y(0). The hard thing on this bank became an easy thing on that bank.

Two things deserve a second look. First, differentiation has turned into multiplication by s — that is the whole calculus-into-algebra promise, made concrete in one line. Second, and this is the part that sets the method apart, the number y(0) shows up uninvited. You did not bolt it on; the rule itself reached into your problem and grabbed the initial value. That is what people mean when they say the Laplace method has the initial conditions built in — they ride along inside the algebra from the first step rather than waiting at the end.

Where y(0) comes from

The rule is not a spell — it falls out of a single integration by parts, and seeing that once will fix it in memory forever. By definition L{y'} is the integral from 0 to infinity of e^(-s t) * y'(t) dt. Integration by parts trades the derivative off y and onto the weight e^(-s t): you differentiate the e^(-s t) (giving a factor -s) and integrate the y'. Out drops a boundary term [e^(-s t) * y(t)] evaluated from 0 to infinity, plus s times the integral of e^(-s t) * y(t) — and that leftover integral is exactly s * Y(s).

Now evaluate that boundary term, and watch the initial condition appear. At the top end, t going to infinity, the weight e^(-s t) crushes y(t) down to zero — this is exactly the exponential-order assumption earning its keep, since it guarantees y cannot outrun the decaying exponential. At the bottom end, t = 0, the weight is e^(0) = 1, so the term is simply y(0). The boundary term is therefore 0 - y(0) = -y(0). Assemble the pieces: L{y'} = -y(0) + s*Y(s). There is the rule, born from one honest calculation, with y(0) arriving as the lower limit of an integral.

Climbing to higher derivatives

A second-order ODE needs the transform of y'', and you get it by applying the first-derivative rule twice — no new idea, just the same move bootstrapped. Think of y'' as the derivative of the function (y'). The rule says L{(y')'} = s * L{y'} - y'(0). Now substitute what you already know for L{y'}, namely s*Y - y(0), and expand. The result is L{y''} = s^2 * Y - s * y(0) - y'(0). Notice the pattern forming: a power of s for the derivative order, and one initial value peeled off per order — this time both y(0) and y'(0).

L{y'}    = s*Y - y(0)
L{y''}   = s^2*Y - s*y(0) - y'(0)
L{y'''}  = s^3*Y - s^2*y(0) - s*y'(0) - y''(0)

General:
L{y^(n)} = s^n*Y - s^(n-1)*y(0) - s^(n-2)*y'(0) - ... - y^(n-1)(0)

Mnemonic: powers of s count DOWN from n-1 ;
          initial values count UP from y(0) to y^(n-1)(0).
The derivative rule and its ladder. Each order contributes one power of s and consumes exactly one initial value.

The general formula L{y^(n)} = s^n * Y - s^(n-1) * y(0) - ... - y^(n-1)(0) is worth reading as bookkeeping rather than memorising. An nth-order initial value problem needs exactly n initial values — y(0) through y^(n-1)(0) — to have one definite answer, and the formula consumes precisely those n numbers, no more, no fewer. That is not a coincidence; it is the same count showing up on both sides. The transform is quietly accounting for every degree of freedom the equation has.

Turning a whole equation into algebra

Here is the rule doing its real job. Take a concrete IVP: y'' + 3 y' + 2 y = 0 with y(0) = 1 and y'(0) = 0. Transform every term, leaning on the linearity you met in guide one so you can go piece by piece. The y'' becomes s^2 Y - s*y(0) - y'(0); the 3 y' becomes 3(sY - y(0)); the 2 y becomes 2Y; and the zero on the right stays zero. Every derivative has vanished, replaced by powers of s and the known numbers y(0) and y'(0).

Substitute the actual values y(0) = 1, y'(0) = 0 and collect. The transformed equation reads (s^2 + 3s + 2) Y - (s + 3) = 0, an ordinary algebraic equation in the single unknown Y. No derivatives, no integrals — just a linear equation you solve by rearranging: Y = (s + 3) / (s^2 + 3s + 2). The differential equation has been transferred entirely into algebra, initial conditions and all. The bracket s^2 + 3s + 2 is, satisfyingly, the very same characteristic polynomial you would have written by guessing y = e^(rt) — the two methods are looking at one structure from two sides.

The mirror rule, and where to be careful

There is a companion worth knowing, because integration shows up as often as differentiation in physics. The transform of an integral is the mirror image of the derivative rule: where differentiating multiplied by s, integrating divides by s. Concretely, the transform of the running integral of y from 0 to t equals Y(s)/s. The symmetry is no accident — integration and differentiation are inverse operations in t, so they become multiplication and division by s in the s-domain. Charge accumulating on a capacitor, distance piling up from velocity: division by s is how the s-domain says 'accumulate'.

Two honest cautions before you trust the derivative rule blindly. First, it is a one-sided transform: the integral starts at 0, so everything is anchored to behaviour from time zero onward, and the y(0) it carries is the value at that left edge — there is no symmetric term from the other end because the upper limit decays away. Second, the rule assumes y is differentiable in the ordinary sense between jumps; if y itself jumps at some interior time, the plain formula misses a contribution there, and you handle that case with the step-function machinery of a later guide rather than this bare rule.

Step back and see what you now hold. One short rule, L{y'} = sY - y(0), simultaneously kills the derivatives and feeds in the initial data, and its higher-order ladder does the same for any order you meet. Apply it across an equation and the whole IVP collapses to algebra in Y(s) — the heart of the master algorithm. You have built steps one and two of that four-step plan. What remains is the return trip: splitting Y(s) and reading y(t) back out, which the next guide hands you in full.