Engineering Mathematics and Numerical Analysis — ESE

Weightage: Engineering Mathematics and Numerical Analysis is one block of ESE Prelims Paper I (General Studies and Engineering Aptitude, 200 marks). The syllabus says each Paper I topic can carry roughly 5 to 15 percent of the paper, and this block usually supplies a handful of short numerical questions. The same tools return in every technical paper, so time spent here pays twice.

1. What the examiner wants

Paper I is common to Civil, Mechanical, Electrical and E&T candidates, so its mathematics is the shared core. Questions are one-step or two-step numerical problems with four options, set under a 1/3 negative-marking rule.

The pattern is consistent. A question names a method and gives small numbers, and the answer is a single value you can reach in a minute. You are rewarded for knowing the formula cold and for estimating when the options are far apart.

2. Matrices and eigenvalues

For a square matrix $A$ of order $n$, the eigenvalues solve $\det (A-\lambda I)=0$.

Three identities answer most questions without solving the equation:

• The sum of the eigenvalues equals the trace of $A$.
• The product of the eigenvalues equals $\det A$.
• Eigenvalues of $A^{-1}$ are $1/\lambda$, of $A^2$ are ${\lambda }^2$, and of $A^T$ are the same as those of $A$.

A matrix is singular exactly when one eigenvalue is zero. A symmetric real matrix has real eigenvalues and orthogonal eigenvectors. For a triangular matrix the eigenvalues are the diagonal entries.

The rank is the number of independent rows. The system $Ax=b$ is consistent when rank of $A$ equals rank of the augmented matrix, and has a unique solution when that rank equals the number of unknowns. The Cayley-Hamilton theorem says $A$ satisfies its own characteristic equation, which lets you write $A^{-1}$ as a polynomial in $A$.

Worked example. For $A=\left(\begin{array}{cc}2& 1\\ 1& 2\end{array}\right)$ the trace is 4 and the determinant is 3. Two numbers with sum 4 and product 3 are 1 and 3, so the eigenvalues are 1 and 3, found without expanding the determinant.

3. Ordinary differential equations

Learn to classify before you solve.

For a first-order equation $\frac{dy}{dx}+Py=Q$, the integrating factor is $e^{\int Pdx}$ and the solution is $y{e}^{\int Pdx}=\int Q{e}^{\int Pdx}dx+C$.

For a second-order linear equation with constant coefficients, $a{y}^{\prime \prime }+b{y}^{\prime }+cy=0$, form the auxiliary equation $a{m}^2+bm+c=0$. The roots decide the form:

This table is also the whole of mechanical vibration and RLC circuit behaviour: overdamped, critically damped and underdamped responses are the three rows.

Worked example. Solve $y^{\prime \prime }+4y^{\prime }+4y=0$. The auxiliary equation $m^2+4m+4=(m+2{)}^2$ has equal roots $-2$, so $y=(C_1+C_{2}x)e^{-2x}$.

4. Laplace transforms

The Laplace transform turns a differential equation into algebra. A short table covers almost every question.

Two rules are used constantly. First shifting: multiplying by $e^{at}$ replaces $s$ by $s-a$. Derivative rule: $\mathcal{L}\{f^{\prime }(t)\}=sF(s)-f(0)$. The final value theorem gives ${\lim }_{t\to \infty }f(t)={\lim }_{s\to 0}sF(s)$, valid only if all poles of $sF(s)$ lie in the left half-plane.

5. Fourier series

A periodic function of period $2\pi$ is written as $a_0/2+\sum (a_n\cos nx+b_n\sin nx)$. Symmetry saves work:

• An even function has only cosine terms, so every $b_n=0$.
• An odd function has only sine terms, so every $a_n=0$.
• A function with half-wave symmetry has only odd harmonics.

At a point of discontinuity the series converges to the average of the left and right limits.

6. Probability and statistics

Keep the standard results ready.

• For events, $P(A\cup B)=P(A)+P(B)-P(A\cap B)$.
• Bayes: $P(A\mid B)=\frac{P(B\mid A)P(A)}{P(B)}$.
• Binomial: $n$ trials with success probability $p$ give mean $np$ and variance $np(1-p)$.
• Poisson: mean and variance are both $\lambda$.
• Normal: about 68, 95 and 99.7 percent of values lie within 1, 2 and 3 standard deviations of the mean.

Engineering use is direct: defect counts follow the Poisson law, test-sample means follow the normal law, and reliability work uses the exponential law with $R(t)=e^{-\lambda t}$.

7. Numerical analysis

When no closed form exists, engineers iterate. Know each method's formula and its order.

Newton-Raphson finds a root of $f(x)=0$ by $x_{n+1}=x_n-f(x_n)/f^{\prime }(x_n)$. It converges quadratically near a simple root, but fails where $f^{\prime }$ is near zero.

Bisection halves an interval that brackets a sign change. It always converges but only linearly, gaining one binary digit per step.

Numerical integration with $n$ equal strips of width $h$:

Simpson's 1/3 rule is exact for polynomials up to degree 3.

Differential equations. For $y^{\prime }=f(x,y)$ with step $h$, Euler's method is $y_{n+1}=y_n+hf(x_n,y_n)$. It is first order. The fourth-order Runge-Kutta method uses four slope estimates per step and is far more accurate.

Worked example. For $y^{\prime }=x+y$, $y(0)=1$, $h=0.1$, Euler gives $y_1=1+0.1(0+1)=1.1$.

8. Using the tools under time pressure

Check the options before computing. If the choices differ by a factor of 10, a rough estimate decides. If they differ in the last digit, compute carefully. A wrong guess costs one-third of a mark, so eliminate two options before guessing.

Common traps

• Using the final value theorem on an unstable system. Check the poles first.
• Applying Simpson's 1/3 rule with an odd number of strips. It needs an even number.
• Forgetting that Newton-Raphson can diverge when the starting point is near a flat region of $f$.
• Treating a Fourier discontinuity value as the function value. It is the midpoint of the jump.
• Mixing up variance and standard deviation in binomial questions.

Memory aids

• "Trace is sum, determinant is product": eigenvalue shortcut.
• "Even means cosine, odd means sine": Fourier symmetry.
• "Quadratic, linear, first order": Newton, bisection, Euler.

Summary

Paper I mathematics is a toolkit. Eigenvalue identities, the three cases of a constant-coefficient second-order equation, a short Laplace table, Fourier symmetry, the standard distributions and the numerical formulas answer almost every question.

Numerical methods matter most: know the formula, the order of accuracy and the condition each one needs.

Exam protocol

• Scan the options for spread before you compute.
• Use identities such as trace and determinant before expanding anything.
• Check conditions (even strips, stable poles) before applying a rule.
• Attempt a question only if you can remove two options.