initial-value problem
A differential equation alone is like a law of motion without a starting point: it says how things change but not where they begin. An initial-value problem supplies the missing starting point. It is the equation together with a condition that fixes the solution's value at one specific instant — answering not just 'how does it evolve?' but 'and where does it start?'.
Concretely, a first-order IVP is the pair dy/dx = f(x, y) together with an initial condition y(x_0) = y_0. The differential equation by itself has a whole family of solution curves filling the plane; the initial condition selects the single curve that passes through the chosen point (x_0, y_0). It is exactly the one extra fact a first-order equation needs (because its order is one), turning a family with one free constant into a definite, particular solution. The name 'initial' comes from the typical physics setup where x is time and y_0 is the state at time x_0 — but mathematically it just means a condition imposed at a single point.
IVPs are the everyday currency of applied mathematics: a projectile launched from a known height with a known speed, a circuit switched on with a known charge, a population census at one year. Crucially, whether an IVP even has a solution, and only one, is not automatic — that is precisely the question the Picard-Lindelof existence-and-uniqueness theorem answers. Numerical solvers like Euler's method are also built specifically to march an IVP forward from its known start.
dy/dx = y with y(0) = 3. The equation's general solution is y = C*e^x; the initial condition forces C = 3, so the IVP has the single answer y = 3*e^x — one curve out of the infinite family, the one passing through (0, 3).
The initial condition collapses a one-parameter family down to one particular solution.
An IVP is not the same as a boundary-value problem. An IVP pins all its data at one point; a BVP imposes conditions at two different points (e.g. y at both ends of a rod). They behave very differently — a BVP can have no solution or infinitely many even when the equation is perfectly tame.