Numerical Analysis and Its Importance
Definition
Numerical Analysis is a branch of mathematics and computer science that focuses on
developing and applying numerical methods (algorithms) to obtain approximate
solutions to mathematical problems that are often too complex or impossible to solve
exactly. Since many real-world problems cannot be solved analytically (with exact
formulas), numerical analysis provides practical tools to handle them with computers.
Significance
Engineering - The finite element method (FEM) allows engineers to predict stresses and
deformations in bridges, buildings, and aircraft.
Physical Sciences - Helps simulate complex physical systems like planetary motion,
quantum mechanics, or weather prediction.
Analytical and Numerical Methods
Analytical methods involve finding exact solutions using mathematical formulas, algebraic
manipulation, and calculus. They are precise and elegant but often limited to problems
that are relatively simple or have well-defined closed-form solutions, such as solving
quadratic equations or integrating basic functions. In contrast, numerical methods rely on
computational algorithms to provide approximate solutions, making them especially
valuable for complex, nonlinear, or large-scale problems that cannot be solved analytically,
such as weather prediction, fluid dynamics, or engineering simulations.
Numerical Errors
Round-off Error
•
Definition: Occurs because computers have finite precision; they can only store a
limited number of decimal or binary digits.
•
Example:
In most computers, storing π\piπ gives 3.14159265353.14159265353.1415926535
instead of the exact infinite value.
Propagation (Accumulated) Error
•
Definition: Errors introduced in early steps of a computation that build up as
calculations proceed.
•
Example:
If each step of an iterative method (like Newton-Raphson or solving ODEs
numerically) has a small error, after many iterations, these errors accumulate and
may significantly affect the final result.
Approximation
The concept of approximation in numerical analysis refers to the process of finding values
that are close enough to the exact solution of a problem when the exact solution is difficult,
impossible, or unnecessary to obtain.
In numerical analysis, approximation plays a central role because it enables us to
transform infinite or highly complex processes into manageable ones. For example,
functions can be approximated using polynomials (Taylor or interpolation polynomials),
integrals can be approximated using numerical integration rules (Trapezoidal or
Simpson’s Rule), and solutions to differential equations can be approximated through
iterative methods (Euler’s method, Runge-Kutta methods).
Significant Figures
The concept of significant figures is important in numerical calculations because it
reflects the precision of measured or computed values and ensures that results are not
reported with misleading accuracy. In practical terms, significant figures indicate how
many digits in a number are meaningful, based on the reliability of the measurement or the
limitations of the computational method.
Convergence and Accuracy
Convergence
Convergence refers to the property of a numerical method where the approximate solution
gets closer to the exact (true) solution as the number of iterations increases or as the step
size decreases.
Accuracy
Accuracy refers to how close a computed or approximate result is to the exact or true
value. It is a measure of the correctness of the solution.
Algorithms
In numerical analysis, algorithms are of central importance because they provide the
step-by-step procedures used to approximate solutions to mathematical problems that
often cannot be solved exactly.
A well-known example is the Newton-Raphson method (also called the Newton method),
which is widely used to find the roots of nonlinear equations f(x)=0. This method is powerful
because it converges very quickly when the initial guess is close to the true root, making it a
standard tool in scientific and engineering computations.
Advantages and Limitations of Numerical Methods
Analytical methods are elegant and provide exact solutions, but they are limited to
simpler, well-defined problems. Numerical methods, on the other hand, are more versatile
and powerful for real-world, complex, and large-scale problems, though they introduce
approximations and errors. Together, both approaches complement each other—
analytical methods provide theoretical insight, while numerical methods make practical
computation possible.
Choice of Step Size
Effect on Accuracy
•
Smaller step size → Higher accuracy
A smaller step size reduces truncation error because the method better
approximates the true curve or function.
•
Too small step size → Round-off error
However, if the step size is extremely small, round-off errors accumulate due to the
limited precision of computers.
Effect on Efficiency
•
Smaller step size → More computations
Reducing the step size increases the number of iterations required, which raises
computational cost
•
Larger step size → Faster but less accurate
A large step size reduces computational work but increases truncation error,
sometimes making the method unreliable.
Real-Life Example
Weather Forecasting
Weather systems are governed by highly complex, nonlinear differential equations that
describe atmospheric dynamics, thermodynamics, and fluid motion (known as the NavierStokes equations). Solving these equations analytically is impossible due to their
complexity and the chaotic nature of the atmosphere. Instead, meteorologists use
numerical weather prediction (NWP) models, which rely on numerical analysis to
approximate solutions.