COL 351 : Analysis and Design of Algorithms

Semester I, 2021-22, CSE, IIT Delhi

Assignment - 3 (due on 24th October, 11:00 PM)

Important Guidelines:

• Each assignment must be done in a group of size at most two.

• Handwritten submissions will not be accepted. Solutions must be typed-up (in Latex, Mi-
crosoft Word, etc.), and submitted in pdf format. Each solution must start on a new page.

• Your answer to each question must be formal and have a proper correctness proof.
No marks will be granted for vague answers with intuition or for algorithms without proof.
You must be very rigorous in providing mathematical detail in support of your arguments.

• Cheating of any form will lead to strict penalty.

1
1 Convex Hull
The Convex Hull of a set P of n points in x-y plane is a minimum subset Q of points in P such that
all points in P can be generated by a convex combination of points in Q. In other words, the points
in Q are corners of the convex-polygon of smallest area that encloses all the points in P .
Design an O(n log n) time Divide-and-Conquer algorithm to compute the convex hull of a set
P of n input points {(x1 , y1 ), (x2 , y2 ), . . . , (xn , yn )}. [15 marks]

Hint:

Step 1: Split P into sets P1 and P2 of size dn/2e by diving along median of x-coordinate of all the points.

Step 2: Recursively compute convex hull of sets P1 and P2 .

Step 3: Finally combine convex hull of P1 and P2 in linear time to obtain convex hull of P .

2
2 Particle Interaction
Some physicists are working on interactions among large numbers of very small charged particles.
Basically, their set-up works as follows. They have an inert lattice structure, and they use this for
placing charged particles at regular spacing along a straight line. Thus we can model their structure
as consisting of the points {1, 2, 3, ..., n} on the real line; and at each of these points j, they have a
particle with charge qj . (Each charge can be either positive or negative.)
They want to study the total force on each particle, by measuring it and then comparing it to a
computational prediction. This computational part is where they need your help. The total net force
on particle j, by Coulomb’s Law, is equal to
X C qi qj X C qi qj
Fj = − .
i<j
(j − i)2 i>j
(j − i)2

They have written the following simple program to compute Fj , for all j.

1 for j = 1, 2, ..., n do
2 Initialize Fj to 0;
3 for i = 1, 2, ..., n do
4 if i < j then
C qi qj
5 Add (j−i) 2 to Fj ;

6 else if i > j then

C qi qj
7 Add − (j−i) 2 to Fj ;

8 end
9 end
10 Output Fj ;
11 end

The running time of this program is O(n2 ). Your task is to design an algorithm that computes
all the forces Fj in O(n log n) time. [15 marks]

3 Distance computation using Matrix Multiplication

Let G = (V, E) be an unweighted undirected graph, and H = (V, EH ) be a undirected graph
obtained from G that satisfy: (x, y) ∈ EH if and only if (x, y) ∈ E or there exists a w ∈ V
such that (x, w), (w, y) ∈ E. Further, let DG denote the distance-matrix of G, and DH be the
distance-matrix of H.

(a) Prove that the graph H = (V, EH ) can be computed from G in O(nω ) time, where ω is the
exponent of matrix-multiplication. [2 marks]
l D (x, y) m
G
(b) Argue that for any x, y ∈ V , DH (x, y) = . [2 marks]
2

3
 (c) Let AG be adjacency matrix of G, and M = DH ∗ AG . Prove that for any x, y ∈ V , the
following holds. [8 marks]
(
2DH (x, y) M (x, y) ≥ degreeG (y) · DH (x, y)
DG (x, y) =
2DH (x, y) − 1 M (x, y) degreeG (y) · DH (x, y)

(d) Use (c) to argue that DG is computable from DH in O(nω ) time. [1 marks]

(e) Prove that all-pairs-distances in n-vertex unweighted undirected graph can be computed in
O(nω log n) time, if ω is larger than two. [2 marks]

4 Universal Hashing
Let U = [0, M − 1] be a universe of M elements, p be a prime number in range [M, 2M ], and
n(<< M ) be an integer. Consider the following hash functions (where r ∈ [1, p − 1]):

H(x) := (x) mod n

Hr (x) := ((rx) mod p) mod n

(a) Compute a random set S as follows:

− Intialize S to empty.
− Repeat n times: choose a random integer in U and add it to S.
Prove the following. [7 marks]
1
P rob(max-chain-length in hash table of S under hash-function H(·) > log2 n) 6
n

(b) Prove that for any given r ∈ [1, p − 1], there exists at least M/n Cn subsets of U of size n in
which maximum chain length in hash-table corresponding to Hr (x) is Θ(n). [3 marks]

(c) Implement H() and Hr () in Python/Java for M = 104 and the following different choices of
sets of size n = 100: For k ∈ [1, n], Sk is union of {0, n, 2n, 3n, . . . , (k − 1)n} and n − k
random elements in U .
Obtain a plot of Max-chain-length for hash functions H(), Hr () over different choices of sets
Sk defined above. Note that you must choose a different random r for each choice of Sk .
Provide a justification for your plots. [5 marks]