Mathwords logoMathwords

Numerical Solution — Definition, Formula & Examples

A numerical solution is an approximate answer to a mathematical problem obtained through computational algorithms rather than exact algebraic manipulation. Instead of producing a formula, numerical methods generate a number (or set of numbers) that is close to the true answer within a specified tolerance.

A numerical solution to an equation f(x)=0f(x) = 0 or a differential equation is a discrete approximation x~\tilde{x} computed by an iterative or finite-step algorithm such that ∣f(x~)∣<ε|f(\tilde{x})| < \varepsilon for a prescribed error tolerance ε>0\varepsilon > 0. The result is a floating-point value rather than a closed-form expression.

How It Works

You start by choosing an algorithm suited to the problem—Newton's method for root-finding, Euler's method for ODEs, or the trapezoidal rule for integrals, for example. The algorithm produces successive approximations, each typically closer to the true value than the last. You stop iterating once the result meets your desired accuracy, measured by how small the residual or the change between iterations has become. Because these methods rely on arithmetic operations, they are implemented on computers and can handle equations that have no known analytical solution.

Worked Example

Problem: Find a numerical solution to x3−x−1=0x^3 - x - 1 = 0 near x=1x = 1 using one step of Newton's method.
Set up Newton's formula: Newton's method updates an initial guess x0x_0 using the formula:
x1=x0−f(x0)f′(x0)x_1 = x_0 - \frac{f(x_0)}{f'(x_0)}
Evaluate at $x_0 = 1$: With f(x)=x3−x−1f(x) = x^3 - x - 1 and f′(x)=3x2−1f'(x) = 3x^2 - 1, compute f(1)=1−1−1=−1f(1) = 1 - 1 - 1 = -1 and f′(1)=3−1=2f'(1) = 3 - 1 = 2.
x1=1−−12=1.5x_1 = 1 - \frac{-1}{2} = 1.5
Check the residual: Evaluate f(1.5)=3.375−1.5−1=0.875f(1.5) = 3.375 - 1.5 - 1 = 0.875. The residual is smaller than ∣f(1)∣=1|f(1)| = 1, so the approximation is improving. Further iterations would bring it closer to the true root x≈1.3247x \approx 1.3247.
Answer: After one Newton step, the numerical solution is approximately x≈1.5x \approx 1.5.

Why It Matters

Most real-world equations in engineering, physics, and finance cannot be solved exactly. Numerical solutions let you solve turbulent-flow equations, price complex financial derivatives, and simulate structural loads—problems where no closed-form answer exists. Courses in numerical analysis and scientific computing are built around these techniques.

Common Mistakes

Mistake: Treating a numerical solution as exact.
Correction: Every numerical solution carries a finite approximation error. Always report the tolerance or number of iterations used, and be aware of round-off and truncation errors that accumulate during computation.

Related Terms