Numerical Analysis I is where mathematics meets the reality of computation. In the “perfect” world of Calculus, we solve for $x$ exactly. In the real world of engineering and data science, equations are often too complex for an exact pen-and-paper solution. Numerical Analysis provides the tools to find “good enough” answers—approximations that are accurate to a specific number of decimal places.
Below is the exam paper download link
PDF Past Paper On Numerical Analysis I For Revision
Above is the exam paper download link
Revision for this unit is a unique challenge. You aren’t just looking for a final number; you are analyzing the process of getting there. How fast does the algorithm work? How much error is introduced at each step? To help you prepare for your finals, we’ve structured this Q&A guide around the core algorithms you’ll encounter in your past paper.
Why Numerical Analysis is the Heart of Modern Engineering
Every time a weather model predicts a storm or an architect simulates the stress on a bridge, they are using Numerical Analysis. Computers cannot handle “infinity” or “infinitesimals” directly; they handle discrete steps. Mastering these methods is essential for anyone looking to work in scientific computing, finance, or machine learning.
Key Revision Questions and Answers
Q1: What is the difference between “Absolute Error” and “Relative Error”?
A: Absolute error is simply the difference between the true value and your approximation ($|x – \hat{x}|$). However, it doesn’t tell the whole story. If you are off by 1 meter when measuring a 10-meter room, that’s a big deal. If you are off by 1 meter when measuring the distance to the moon, it’s negligible. Relative Error scales the absolute error by the true value, giving you a percentage that reflects the actual significance of the mistake.
Q2: When should I use the Bisection Method versus the Newton-Raphson Method?
A: The Bisection Method is the “slow and steady” approach. It is guaranteed to find a root as long as the function changes sign over an interval, but it takes many steps. Newton-Raphson is the “sprinter.” It uses the derivative to zoom toward the root incredibly fast. However, it can fail if your starting guess is bad or if the derivative is zero. In exams, look for clues about “convergence speed” to decide which one to discuss.
Q3: What is “Interpolation,” and how does it differ from Curve Fitting?
A: In Interpolation (like Lagrange or Newton polynomials), you are creating a function that passes exactly through every data point you have. It’s used to find values between known points. Curve Fitting (like Least Squares) creates a trend line that follows the general shape of the data but might not actually touch any of the individual points.
Q4: Explain the concept of “Round-off Error” in computer calculations.
A: Computers have finite memory. They can only store a certain number of digits for any given number (floating-point representation). When you perform millions of calculations, these tiny chopped-off decimals add up. This is known as Round-off Error. A key part of your revision should be understanding “stability”—identifying which algorithms keep these errors small and which ones let them explode.
Q5: What is the goal of “Numerical Integration” (e.g., Simpson’s Rule)?
A: Sometimes, an integral is impossible to solve analytically (like the bell curve in statistics). Numerical integration rules, like the Trapezoidal Rule or Simpson’s Rule, break the area under the curve into simple shapes—trapezoids or parabolas—and sum their areas. Simpson’s Rule is generally more accurate because it uses curved tops to better match the original function.
How to Use This Past Paper for Your Revision
Numerical Analysis is a subject that requires a calculator and patience. Once you Download PDF Past Paper On Numerical Analysis I For Revision, don’t just stop at the first iteration. Most exam questions will ask you to “Perform three iterations” or “Find the root correct to 4 decimal places.”
Practice the “table method” for your iterations. Keeping your work organized in columns for $n$, $x_n$, and $f(x_n)$ is the best way to avoid the small arithmetic errors that can derail an entire problem. By working through these real-world exam scenarios, you will develop the precision needed to master the art of approximation.
Last updated on: March 21, 2026