01 Mar 2026

Numerical Methods Syllabus Map

A compact SEM-I map for numerical methods from roots of equations to finite difference methods.

msc semester-i numerical-methods computational-techniques

Numerical methods begin with a practical difficulty: many equations in physics and mathematics cannot be solved exactly in a simple closed form. Even when an exact solution exists, it may be too long for repeated calculation. The numerical approach is to replace the original problem by a sequence of clear arithmetic steps.

Thus the question is not only “what is the answer?” but also “how close is the approximate answer to the true one?” A good numerical solution must therefore include the method, the approximation used, and a check on the result.

Flow of ideas

1. Roots of functions locate where $f(x)=0$.
2. Iteration methods improve an approximate answer step by step.
3. Gauss elimination solves simultaneous linear equations.
4. Eigenvalues and eigenvectors describe matrix modes and characteristic values.
5. Interpolation and extrapolation estimate values inside or outside tabulated data.
6. Curve fitting and least squares find the best approximate relation from data.
7. Numerical integration estimates areas and integrals by trapezoid and Simpson rules.
8. Runge-Kutta method solves first-order differential equations numerically.
9. Finite difference method replaces derivatives by differences on a grid.

Basic numerical philosophy

A numerical answer is not only a number. It must be checked for convergence, stability, step-size dependence, and physical reasonableness.

The usual path is:

1. start from the mathematical problem;
2. choose a numerical formula suited to that problem;
3. compute the result step by step;
4. estimate or reduce the error;
5. interpret the answer in the original physical setting.

Common sources of error

• Round-off error comes from finite precision arithmetic.
• Truncation error comes from cutting off an infinite or exact process.
• Iteration error comes from stopping an iterative method before the exact limit.
• Modelling error comes from the assumptions used before computation starts.

In an examination answer

In an examination answer, first state the method and the formula being used. Then show the substitution of numerical values clearly. A numerical result should be rounded only at the end, because early rounding may change the final answer.

For a complete answer, also mention the check used:

• in root finding, substitute the answer in $f(x)$;
• in integration, compare the answer with a smaller step size if possible;
• in iteration, show that successive approximations are becoming closer;
• in linear equations, substitute the solution back into the original equations.

Practice questions

1. Distinguish between round-off error and truncation error.
2. Explain why convergence is important in an iterative method.
3. State one advantage and one limitation of numerical methods.
4. Why should a numerical result be checked by substitution or step-size reduction?
5. Arrange the following methods in a suitable study order: interpolation, root finding, numerical integration, and differential equations.