An approximation is a representation of a value, quantity, function, or object that is intentionally not exact but is sufficiently close for a particular purpose. People use approximations when the exact result is unknown, unavailable, too costly to obtain, or more detailed than necessary. In written or mathematical contexts the phrase "approximately equal" is often shown with the sign ≈, and the same idea can be expressed with words such as "about," "around," or "roughly." For a concise definition see introductory references; in practice the acceptable difference between an approximation and the exact value is set by the intended tolerance.

Characteristics and types

Approximations vary by intended accuracy and method. Common categories include rounding (replacing a number with a nearby simpler value), truncation (cutting off digits or terms), numerical approximation (computing values by algorithms), and analytic approximation (using series, asymptotic expansions, or simpler functions to represent a complex one). Each approach balances three factors: simplicity, accuracy, and cost (time or computational resources). An approximation has an associated error or residual that measures how far it is from the exact value; controlling that error is central to choosing or designing a method.

Typical notations and conventions help communicate approximation quality. The symbol ≈ indicates approximate equality, while a tilde ~ sometimes denotes asymptotic equivalence. Numerical contexts often display a value together with an accuracy indicator, for example 3.14 for π (more precisely π ≈ 3.1416) or monetary amounts rounded to the nearest cent. Visual aids or plotted comparisons are frequently used to show how well an approximation matches an exact curve or dataset.

History and notable methods

Historically, approximation has been crucial for mathematics, astronomy, engineering, and navigation. A famous early example is Archimedes' method of inscribed and circumscribed polygons to bound the circumference of a circle and thereby obtain estimates for π; he used polygons with many sides to tighten upper and lower bounds, an approach that anticipates modern limit and convergence ideas. See more about Archimedes and his work with polygons around a circle. Analytical techniques such as Taylor series, Fourier series, continued fractions, and perturbation methods provide powerful ways to approximate functions.

Practical methods and examples

  • Rounding and truncation: everyday simplifications like saying "about 80 dollars" instead of $78.23.
  • Discrete numerical methods: finite differences or finite element discretization for differential equations.
  • Statistical and Monte Carlo approximations: sampling-based estimates when deterministic calculation is impractical.
  • Series and asymptotics: using a few terms of an expansion to approximate functions for large or small parameters.

Simple arithmetic approximations include adding 2.0002 + 2.0003 by treating both as 2 + 2 to get an initial rough sum, then refining as needed. In many engineering applications, an approximate model that captures dominant effects is preferred to an exact but intractable description.

Importance, error, and distinctions

Understanding approximation error is essential: absolute error quantifies the numerical gap, while relative error scales that gap by the magnitude of the exact value. Tolerance requirements vary by field — scientific computation may demand extremely small relative errors, while rough estimates suffice for planning or communication. It is also useful to distinguish approximation from estimation: "approximation" often refers to replacing a mathematical object with a simpler one, whereas "estimation" frequently denotes inferring an unknown parameter from data. For further reading on methods and examples, consult resources about π and numerical techniques; for instance, introductory material on the calculation of π illustrates many approximation ideas.