4 7

Home Answer Notifications Search Quora Add Question or Link

What must be my strategy to score full marks

in the GATE mathematics (also suggest best
book for practicing prev. year’s questions)?
Sreenath Vemula, AIR 82 in GATE
Answered Sep 16, 2017

Mathematics is a very scoring subject. And it is very important for those who just
want to pass or to get below 2000 or top 100. Who want to miss easy 10+ marks???

Some of the problems students face related to mathematics are:

Students have studied it in the first year and have completely forgotten

Mathematics is widely regarded as one of the toughest subjects

Students think they can’t learn that subject since it is beyond them

So overcome your fears and see that it is very easy provided, you study well. Did I tell
you, my most favorite subject is mathematics. I learnt mathematics by practice only!
When I see a new topic, I too feel blank inside my head. Then I realize that all I need
to do is to understand the concepts by practicing then I feel comfortable in doing
those concepts.
I will try to explain important topics here and give some tips to prepare mathematics.
Finally I will give some timetable to study it so that you finish that subject. Before we
start, see the GATE syllabus here:

A E , Civil , Chemical , CS , ECE , EE , ME , Petroleum ,

Instrumentation

All branches except CS have almost same syllabus. So I will focus on these common
topics. You can study from either ACE material or Made Easy material or B. S. Gravel
or any good textbook along with previous year papers.

Try to spend daily an hour on Mathematics till you become confident in it.

Linear Algebra: Generally the topics asked are related to finding rank,
determinant, Eigenvalues and Eigenvectors, Solution of simultaneous
equations, Finding inverse and power of matrix.

Rank: It is the minimum of ( independent rows or columns ) in the given

matrix. To find the rank. try to write the given equations in to matrix form
and reduce it into row echelon form . The number of non-zero rows left
in the matrix is the rank of the matrix. Ex: Rank of matrix

Determinant: You can solve these problems in two methods:

Expand the matrix by finding co-factors and then simplifying it gives

the determinant and straightforward. Refer here for complete
procedure: The Determinant of a Square Matrix

Next approach is very simple and easy. Try to convert the given matrix
into row echelon form. Then the determinant is just the product of
diagonal terms. Only point you need to keep in mind is, when you
interchange two adjacent rows, rank doesn’t change but determinant
changes to opposite sign. So if you interchange two rows say ith and jth
. Then, determinant changes the sign to (−1)j−i . Also another point to
remember is not to multiply any row by a constant. Rank doesn’t
change but determinant changes.

Solution of simultaneous equations: To solve there questions, try to

write the Augmented matrix [A|b] from Ax = b. Find the rank of [A|b]
and also for A.
 If R(A) = R([A|b]) = n (no. of rows or columns in the given matrix),
then we have unique solution.

If R(A) = R([A|b]) < n (no. of rows or columns in the given matrix),

then we have multiple solutions.

If R(A) ≠ R([A|b]), then we have no solution.

Eigenvalues and Eigenvectors: Eigenvalues and Eigenvectors are the

solutions the equation Ax = λx, λ is the Eigenvalue and x is the
Eigenvector. To solve these types of problems, since we have one equation
and two unknowns (λ and x ) so there is no unique solution to given
equation. Refer here for a solved example: Eigenvalues and Eigenvectors
example

To find Eigenvalues, write given equation as (A − λI )x = 0. Find the

determinant of (A − λI ). λ is Eigenvalues obtained.

Next for each Eigenvalue λ obtained, find the Eigenvector

corresponding to it. Since the equation becomes to some Bx = 0, we
don’t have a unique solution. Assume one of the values in vector as 1.
Find the rest of the terms.

Finding inverse and Power of matrix: use Cayley Hamilton theorem to

solve these problems. Every matrix satisfies its own characteristic
equation |A − λI | = 0. From this find the equation in terms of λ.
Replace λ with A and then we get characteristic equation. From this, we
can find any power or inverse easily. Refer here for a solved problem:
Cayley Hamilton Theorem Examples

Calculus: Some topics to focus are finding maximum and minimum values,
Finding Limits, Continuity and Differentiability of functions, Taylor and
Maclaurin series, Vector calculus, Finding area and Volumes,

Maximum and Minimum values: Differentiate the given equation and

find f ′ (x) = 0. Then, find f " (x) at the same values obtained. Refer this
for further clarification: Minimum and Maximum Values

f " (x) < 0, then value is local maxima

f " (x) > 0, then value is local minima

f " (x) = 0, then nothing can be said

Also check with the boundaries too. They could be maximum or

minimum in the given interval.

Limit, Continuity and Differentiability: Refer this for some solved

examples: Continuity and Differentiability

Limits: To find the limit of form, 0/0 or ∞/∞ we need to use L-

Hospital rule. 0/∞0 and ∞/0 → ±∞ and so limit does not exist, if
limit is checked from left and right of zero. For 0 × ∞ form, write it as
0/(1/∞) = 0/0 form and use L-Hospital rule. A simple approach to
solve majority of the trigonometric and logarithmic problems is by
expanding the terms in x. And then simplifying gives directly the limit.

Ex. of above limit application:

3 5
sin x − x (x − x /6 + x /120−. . . ) − x
−1
limx→0 = limx→0 =
3 3 6
x x

Continuity: To find whether function is continuous or not, simply find

the limx→0 and limx→0 and f (0) and see that all three values are
+ −

equal.

Differentiability: Along with continuity, the function need to have a

unique limit of differentiated value.
f (x + h) − f (x) f (x + h) − f (x)
′
f (x)| = limh→0 + = limh→0 −
x=0
h h
 For two variable problems, limit has to be same in any order we apply
for continuity and differentaiability. Also if limit doesn’t exist, just
solve the problems with one the given order.

Taylor series and Maclaurin Series: To solve problems related to this,

find the function first, second and third etc. derivatives at x=a (for Taylor)
and x=0 (for Maclaurin. Then get the expansion of the function.

Taylor series is the expansion of function along x=a (from: Taylor Series )

Maclaurin series is the expansion of function along x=0

Definite and Indefinite integrals: Indefinite integrals are very tricky. Try
to solve solutions and check if we get given question. Definite integrals are
− π/2 n π/2 n
the toughest. Learn formulas of ∫ √x , ∫
0
sin x, ∫
0
cos x,

Solve some problems of definite integrals from here: Computing

Definite Integrals

Most of the definite problems can be solved by method of substitutions

2 − 2 2
like x = t for √x , x= a tan x for x + a in denominator,
−−− −−−
x = a sin x for √ a2 − x 2 and so on. Refer here for examples of all
models: Substitution Rule for Indefinite Integrals , More Substitution
Rule
b
Area between two curves is ∫a f (x) − g(x)dx is the area between
f (x) and g(x) where a and b are the two intersection points of the
curve. To find area of the curve and x axis, take g(x) as zero. Refer here
for some solved examples of Area Between Curves

To calculate volumes, use ∫ Ady or ∫ Adx. Where, A represents the

area of the typical disc. Refer here for some solved example: Volume of
Revolution

Vector Calculus: The questions asked from this topic are very easy if we
know the formulas. Very important operator is Del:
∂ → ∂ → ∂ →
∇ = i + j + k
∂x ∂y ∂z

→ → →
Curl F is defined as the cross product between ∇ and F =∇× F

→ →
Divergence or div F is defined as ∇. F . Refer here to solve some
problems: Curl and Divergence

Gradient is defined for a scalar only where as we defined Curl and

Divergence to a vector. Gradient is defined as
 → → →
→ ∂F → ∂F → ∂F →
∇F = i + j + k
∂x ∂y ∂z . The rate of change of function
is defined by gradient.

To get directional derivative of the function, simply calculate the

→
→ →
gradient and calculate ∇ f. u (where . is dot product and u is unit
vector in the direction of given vector. Refer here for some solved
problems: Directional Derivatives . To calculate the maximum rate of
→
→ ∇f
∇ f.
→
change given function f (x, y, z), simply compute | ∇ f|

To compute tangent and normal of a function, compute gradient first.

Refer here for solved problems: Gradient Vector, Tangent Planes and
Normal Lines

Tangent of a function f (x, y, z) at (x0 , y0 , z0 ) is represented by:

Similarly normal is represented by

If the field is conservative , we can calculate potential which is

nothing but the gradient of function.

For line integrals, just take the equation of the path given and integrate
along it. Refer here: Line, Surface and Volume Integrals

Greens Theorem: Refer here for example: Green's Theorem . It

converts line integral to area integral. It is defined only for 2D as
shown:

Stokes Theorem: It is the higher version of Greens Theorem. It converts

surface integral to volume integral. Refer here for example: Stokes'
Theorem . It is defined as:

Gauss Divergence Theorem: It converts volume integral to surface

integral (or reverse of Stokes theorem). Refer here for example: Gauss
Divergence Theorem

Transformations: Key points to focus would be finding of transform for

given function specially half sine, cosine, unit step function, and Dirac Delta
function, and its inverse or convolution and its application to compute the
differential equations.

Fourier Series:

Fourier Series Tutorial I

Fourier Series Tutorial II

Fourier Series Examples

Laplace Transforms:

Laplace Transforms Overview I

Laplace Transforms Overview II

 Step Functions

Initial Value Problems (IVP)

Dirac Delta Function

Formulas sheet: Laplace Transforms Formulas

Inverse Laplace Transforms

Dirac Delta function: Function is infinite at a single value and is zero at

rest of the values. Applying Fourier and Laplace transforms to it very
important. Refer here for its applications

z‐Transform: Key points to note are applying z-transforms, inverse z-

transforms, finding poles and zeros, and application to compute power
series.

Zeros are locations where numerator of function is zero and Pole is a

location where the denominator is zero so function is infinite there. For
z − 1

ex. consider H (z) =

z + 1
. Zeros are z = 1 and poles are z = −1 .

Z-Transforms Complete

Z-Transforms

Z-Transforms Overview

Properties of Z-Transforms

Power series expansion exists only if region of convergence |R| < 1

Power Series and Functions

Ordinary Differential Equations: Some key points to remember is learning

different first order types, for second order with particular integral as
x n
sin x, cos x, e , x , combination of any of these, application of method of
variation of parameters to any general function like tan x, sec x, log x etc.
and application of Euler-Cauchy method for variable coefficients. Solving
IVP, BVP.

Order is the highest order differential equation. Degree is the power of

highest order in given differential equation after differential equation is
radical free (meaning free from 1/3 power etc.)

First Order Linear equations: If only y ′ term is present it is first order.

Linear if y ′ is free linear. Some of the types of standard models available
are:

Linear

Homogeneous

Exact

Variable Separable

Bernoulli's equation

Above types can be found here: Recognizing types of ODE and First
Order DE's and Problems I

Second Order Equations with constant coefficients: Standard form is

′ 2
ay " +by + cy = f (x) = (aD . Where,
+ bD + c)y = f (x)

D = d/dx. To solve these equations,

First form an characteristic equation i,e, aD 2 + bD + c = 0. Solve for

D. Let D = a be a root,
ax
If roots are unique, solution is y = c1 e . By using superposition
combine all the roots. Real & Distinct Roots
ax
If two roots are equal say D = a, then solution is y = (C0 + xC1 )e

.Differential Equations - Repeated Roots

 If roots are complex say D = a ± ib, then the solution is
ax
y = (c0 sin bx + c1 cos bx)e . Differential Equations - Complex
Roots

After finding the roots, we need to particular integral (PI). i,e, solution
which is just satisfying the given function. We have F (D)y = f (x)

1
ax
e
If PI is of form eax , then solution is F (a)

If PI is of form sin ax, cos ax then replace D 2 with −a2 .

If PI is of xax , then simply find partial fractions of F (D) and then

integrate.

If we have some combination of eax , sin bx, then first write sin bx or
cos bx into exponential form using Euler identity
ibx
e = cos bx + i sin bx. Then combine both of the exponential terms
and solve.

There is an alternative way to find solution. If PI is of form eax , the

solution is of form Aeax . If PI is of form sin ax, or cos ax the solution
is of form A cos ax + B sin ax. If PI is of form c0 x2 + c1 x + c2, then
solution is of form ax2 + bx + c. Once we assume the solution, then
substitute in the given differential equation and find the coefficients.

Refer here for solved problems: Problems I and Problems II

Application of Differential Equations

Method of variation of parameters: It is used for any particular integral.

Refer here how to solve: Variation of Parameters I , Solving tan (x)

Variable coefficients: If the differential equation is of form

n n n−1 n−1 r
c0 x D + c1 x D +. . . +cn )y = f (x), we need to assume y = x

as solution and substitute in the differential equation and solve for r.

Refer here for solved examples: Differential Equations - Euler Equations

Partial Differential Equations: variable separable method is widely used in

these problems. Some key points are:

Partial derivatives determination is important to solve these problems.

∂z ∂z
a + b = c
If First order PDE is of form, ∂x ∂y , The solution is solved by
dx dy dz
= =
using a b c
. Refer this to solve some problems Linear PDE

Refer this to solve non linear first order PDE: Problems I and Problems II

General second order equation is

.
au xx + buxy + cu yy + du x + eu y + f u = g(x, y)

If b2 − 4ac < 0, PDE is elliptic. Ex: Laplace equation

2
If b − 4ac = 0, PDE is parabolic. Ex: Heat equation

If b2 − 4ac < 0, PDE is hyperbolic. Ex: Wave equation

2
∂u ∂ u K0
= κ κ =
One-Dimensional diffusion equation: ∂t 2
∂ x
. Where cρ is
called thermal diffusion. Refer here for solved example: 1D Diffusion
equation

Heat equation: α2 uxx = ut .

We get the solution in variable separable form by assuming of

y = X(x)T (t).

Substituting the above yields to α2 X " (x)T (t) = X(x)T (t)

′

′
X " (x) T (t)
⇒ = = C
X(x)
2
α T (t) (Constant. Since LHS is a function of x
and RHS is a function of t).
 Solve the two equations and substitute boundary conditions.

Solved example of Heat equation

Wave equation: α2 uxx = u tt . Solve as above and solve by variable

separable method. Refer here for solved example: Wave Equation

Numerical Methods: Key topics to focus are solution of algebraic equation

by Newton’s method, single step numerical differentiation and Simpson and
Trapezoidal methods.

Simpson’s 1/3 rule: Used to determine area approximately.

b
h
∫ f (x)dx = [y0 + yn + 4 (∑ yi ) + 2 (∑ yi )]
i=odd i=even
Given by: a 3

b−a ¯
¯¯¯
¯¯¯¯
¯ ¯
¯¯¯
¯¯¯¯
¯
Error E = −
180
Δ y,
4
where Δ4 y is fourth forward difference.

Solved example here: Simpson’s 1/3 Rule Example

Trapezoidal rule:
b
y0 +yn n−1
∫ f (x)dx = h [ + ∑ yi ]
i=1
Given by a 2

b−a ¯
¯¯¯
¯¯¯¯
¯
Error E = −
12
2
Δ y

Solved example here: Trapezoidal Rule Example

Simpson’s 3/8 rule:

Given by:
b
3h
∫ f (x)dx = [y0 + yn + 3 (∑ yi ) + 2 (∑ yi )]
8 i=1,2,4,5,⋯ i=3,6,⋯
a

3(b−a) ¯
¯¯¯
¯¯¯¯
¯
4
Error E = −
80
h Δ y
4

Solved example here: Simpson’s 3/8 Rule Example

Euler’s forward method: Refer this for solved example

2
h
Given by: yi+1 = yi + hf (x i , yi ) +
2!
f (x i , yi ) + ⋯

It is a single step forward method. Error order O(h2 )

For Euler’s backward method, use yi−1 instead of yi+1 and h with −h

Range-Kutta 2nd Order: It has error order 2. Given by

h
yk+1 = yk + (f (x k , yk ) + f (x k + h, yk + hf (x k )))
2

Range-Kutta 4th Order: It has error order 4 i.e O(h5 ).

k1 +2k2 +2k3 +k4
Given by yk+1 = yk +
6
, where,

k1 = hf (x k , yk )

h k1
k2 = hf (x k + , yk + )
2 2

h k2
k3 = hf (x k + , yk + )
2 2

k4 = hf (x k + h, yk + k3 )

Refer here for Newton Interpolating Polynomial and Lagrangian

Interpolating Polynomial

Newton-Raphson method: Used for solving roots of algebraic and

f (x i )

transcendental equations. Given by xi+1 = xi + ′

f (x i ) . Refer here for
solved examples

Bisection Method and False Position Method - Linear convergence; Secant

Method - 1.618 and Newton’s Raphsons Method - 2
 Probability and Statistics: Key topics to focus are Bayes Theorem,
Conditional and Joint probability, Binomial, Poisson, normal and
exponential distribution — there mean, variance and standard deviation,
Correlation and regression.

Properties of Probability :

0 ≤ P (A) ≤ 1

′
P (A) = 1 − P (A )

P (∅) = 0

P (∪) = 1

Conditional Probability : Given by

P (A ∩ B) = P (A|B) × P (B)

P (A ∩ B) = P (B|A) × P (A)

P (E i )P (E i |A)
P (Ei |A) = n

∑ P (E k )P (A|E k )
Bayes Theorem: k=0
, where
P (E i ∩A)
P (Ei |A) =
P (A) . Refer here for solved examples

Discrete Distribution :
n x n−x
Binomial Distribution : f (x) = ( )p (1 − p)
x
, denoted by
2
X ∼ b(n, p). Mean μ = np, Variance σ = np(1 − p)

x−1
Geometric Distribution : f (x) = P (X = x) = (1 − p) p.

Cumulative distribution function is

x 1
F (x) = P (X ≤ x) = 1 − (1 − p) . Mean μ = E(X) =
p
, Variance
2 1−p
σ = V ar(X) = 2
p

Negative Binomial Distribution :

x−1 x−r r r
f (x) = P (X = x) = (
r−1
)(1 − p) p , Mean μ = E(X) =
p
and
r(1−p)
2
Variance σ = 2
p

−λ x
e λ
Poisson Distribution : f (x) =
x!
, Mean μ = λ, Variance σ 2 = λ

Continuous Distribution :
1 −x/θ
Exponential Distribution : f (x) =
θ
e , Mean μ = θ, Variance
2 2
σ = θ

−1 x−μ 2
1 ( )
Normal Distribution : f (x) = − e σ

σ√2π
2
, Mean μ and the
variance of X is σ 2

X − μ

Substituting Z =
σ
converts normal distribution into standard
normal distribution. Solve example problems here

Hypothesis testing can be studied from here and Correlation and

Regression from here

For specific doubts, comment below.

It would be grateful if someone could assist me in creating the PDF (with proper links
and formulas)of this file and add a link here so that it would be very helpful to all.

26.7k Views · View Upvoters · View Sharers · Not for Reproduction · Answer requested by Praveen
Kumar Yadav, Sourodeep Bhattacharya (सौरदीप भ ाचाय), and 14 more

You upvoted this

Upvote · 458 Share · 1

Add a comment... Recommended All

Abhishek Roy
Feb 4
One question, how did you write equation in a quora answer. What did you use?
Reply · Upvote · Downvote · Report