Mathematical Foundations of Computer Science Lecture Outline

Mathematical Foundations of Computer Science
Lecture Outline
December 4, 2018
The Probabilistic Method

For any graph G = (V, E), a set of vertices D ⊆ V is called a dominating set if every vertex
in V \ D is adjacent to a vertex in D.
Example. Prove that any connected graph G = (V, E) with n ≥ 2 vertices and minimum
degree δ(G) = δ, contains a dominating set of size at most n(1+log(1+δ))
1+δ .
Solution. For each vertex v ∈ V , add it to the set X independently with probability p.
Let Y ⊆ V \ X be the vertices that are not dominated by X, i.e., they are vertices in V \ X
that are not dominated by X. Then X ∪ Y is a dominating set for G. We will now show
that E[X ∪ Y ] is not too large. Since X and Y are disjoint sets, we have
E[X ∪ Y ] = E[X] + E[Y ] (1)
We consider the following random variables.

Xv : random variable that is 1 if vertex v is in X, 0, otherwise.
Yv : random variable that is 1 if vertex v and all of its neighbors are not in X, 0, otherwise.
X
X= Xv
v
X
∴ E[X] = Pr[Xv = 1]
v
= np
X
Y = Yv
v
X
∴ E[Y ] = Pr[Yv = 1]
v
X
= (1 − p)deg(v)+1
v
X
≤ (1 − p)δ+1
v
= n(1 − p)δ+1
Plugging the values of E[X] and E[Y ] in (1) we get
E[X ∪ Y ] ≤ np + n(1 − p)δ+1 ≤ np + ne−p(δ+1) ,

2 Lecture Outline December 4, 2018
The last expression is minimized when
ln(1 + δ)
p=
1+δ
n(1+ln(1+δ))
Thus, we can find a dominating set of size at most 1+δ .
Recall that a tournament is a directed graph with exactly one directed edge between any
pair of vertices. A tournament G = (V, E) is called k-dominated if for every set of k vertices
v1 , v2 , . . . , vk , there exists another vertex u ∈ V such that (u, vi ) ∈ E, for i = 1, 2, . . . , k.
Example. Prove that if n is large enough then there is a k-dominated tournament on n

vertices.
Solution. Construct a random tournament G in which an edge between any two vertices
u and v is directed towards u with probability 12 and towards v with probability 21 . The bad
event for our random process is that G is not k-dominated. We will calculate the probability
of this bad event as follows. Let S be a fixed set of k vertices in G. The probability that a
vertex u outside of S does not dominate set S is given by 1 − (1/2)k . Thus the probability
that S is not dominated by any of the n − k vertices outside of S is given by (1 − 1/2k )n−k .
n

Since there are k possibilities for set S, the probability of some set of k vertices in G not
being dominated is at most
1 n−k

n
1− k (2)
k 2
If the above expression is less than 1, it means that the probability of the random tourna-
ment G being k-dominated is strictly larger than 0, which means that such a tournament
exists. We will now show that if n/ ln n > k2k then the expression (2) is less than 1.
1 n−k nk k−n

n
1− k ≤ · e 2k (using 1 + x ≤ ex , ∀x ∈ R)
k 2 k!
ek ln n k
−n
= · e 2k 2k (since n = eln n )
k!
k
e 2k k ln n− nk
= ·e 2
k!
e 1
≤ · (since n/ ln n > k2k )
k! e
1
= <1
k!
Note that for large values of k, n > k 2 2k satisfies the inequality n/ ln n > k2k . This is
because when n = k 2 2k , we have
n k 2 2k
=
ln n ln(k 2 2k )
December 4, 2018 Lecture Outline 3
Note that for the last term to be larger than k2k , it must be that
e k
ln(k 2 2) < k ⇒ k 2 2k < ek ⇒ k 2 <
2
which is true for sufficiently large values of k.
Operations on Relations
We can take a relation or a pair of relations and produce a new relation. Since a relation R
from set A to set B is a subset of A × B, operations that apply to sets apply to relations.
Example. Let A = {1, 2, 3} and B = {a, b, c, d}. Let R1 = {(1, a), (1, c), (2, c), (3, a)}.
Let R2 = {(1, b), (1, c), (1, d), (2, b)}. Then we have
R1 ∪ R2 = {(1, a), (1, b), (1, c), (1, d), (2, b), (2, c), (3, a)}
R1 ∩ R2 = {(1, c)}
R1 \ R2 = {(1, a), (2, c), (3, a)}
R2 \ R1 = {(1, b), (1, d), (2, b)}
Example. Let A and B be the set of all students and the set of all courses at a school,
respectively. Suppose R1 consists of all ordered pairs (a, b), where a is a student who has
taken course b, and R2 consists of all ordered pairs (a, b), where a is a student who requires
course b to graduate. What are the relations R1 ∪ R2 , R1 ∩ R2 , R1 ⊕ R2 , R1 \ R2 , and
R2 \ R1 ?
Solution. R1 ∪ R2 consists of all ordered pairs (a, b), where a is a student who has taken
course b or requires course b to graduate.
R1 ∩ R2 consists of all ordered pairs (a, b), where a is a student who has taken course b and
requires course b to graduate.
R1 ⊕ R2 consists of all ordered pairs (a, b), where a is a student who has taken course b or
requires course b to graduate, but not both.
R1 \ R2 consists of all ordered pairs (a, b), where a is a student who has taken course b but
does not require it to graduate.
R2 \ R1 consists of all ordered pairs (a, b), where a is a student who required course b to
graduate but has not taken it.
Inverse Relation
Let R be a relation from A to B. Then the inverse of R, written R−1 , is the relation from
B to A defined by
R−1 = {(b, a) | (a, b) ∈ R}
Example. Let A = {a, b, c} and let R = {(a, a), (a, b), (b, a), (c, a)}. Then
R−1 = {(a, a), (b, a), (a, b), (a, c)}
Note that R and R−1 are almost equal.
4 Lecture Outline December 4, 2018
Example. A relation R on a set A is symmetric iff R = R−1 .
Solution. ( =⇒ ) Suppose R is symmetric on A. We will prove that R = R−1 by showing

that R ⊆ R−1 and R−1 ⊆ R. We will prove R ⊆ R−1 by showing that an arbitrary element
(a, b) ∈ R is also in R−1 . Since R is symmetric, (b, a) ∈ R. By definition of R−1 , since
(b, a) ∈ R, it must be that (a, b) ∈ R−1 . To prove R−1 ⊆ R, we will show that an arbitrary
element (a, b) ∈ R−1 is also in R. By definition of R−1 , it must be that (b, a) ∈ R. Since R
is symmetric, (a, b) must also be in R.
( ⇐= ) Suppose that R = R−1 . Let (a, b) be an arbitrary ordered pair in R. To prove that
R is symmetric we need to show that (b, a) ∈ R. By definition of R−1 , (b, a) ∈ R−1 . Since
R = R−1 , R must contain (b, a).
Composition of Relations
Let R be a relation from A to B and S be a relation from B to C. The composition of S
with R is the relation from A to C:
S ◦ R = {(x, z) | there exists a y ∈ B such that x R y and y S z}
Example. Let A = {1, 2, 3, 4}, B = {3, 4, 5, 6}, and C = {a, b, c}. Let R and S be
relations from A to B and from B to C, respectively, where
R = {(1, 3), (3, 3), (3, 4), (4, 5), (4, 6)}
S = {(3, b), (4, a), (4, c), (5, a), (5, b), (6, c)}
What is the composite of the relations R and S?
Solution. S ◦ R = {(1, b), (3, a), (3, b), (3, c), (4, a), (4, b), (4, c)}
Let R be a relation on a set A. The powers Rn , n = 1, 2, 3, . . . , are defined recursively by
R1 = R and Rn+1 = Rn ◦ R
Observe that R2 = R ◦ R, R3 = R2 ◦ R = (R ◦ R) ◦ R, and so on.
Example. Let R be a relation on a set A. Then R is transitive iff Rn ⊆ R, for all n ≥ 1.
Solution. We first show that if Rn ⊆ R, for all n ≥ 1, then R is transitive. Note that if
(a, b) ∈ R and (b, c) ∈ R then (a, c) ∈ R2 . Since R2 ⊆ R, it must be that (a, c) ∈ R, which
means that R is transitive.
We will prove R is transitive =⇒ Rn ⊆ R, for all n ≥ 1, using induction on n.
Induction hypothesis: Assume that if R is transitive then Rk ⊆ R, for some k ≥ 1.
Base Case: The claim holds tivially when n = 1, since R1 = R.
Induction Step: We want to prove the claim when n = k + 1. In other words, we want to
prove that if R is transitive then Rk+1 ⊆ R. We will prove this by showing that an arbitrary
but particular ordered pair (a, b) in Rk+1 is also present in R. By definition, Rk+1 = Rk ◦R.
December 4, 2018 Lecture Outline 5
Since (a, b) ∈ Rk+1 , there must be a c, such that (a, c) ∈ R and (c, b) ∈ Rk . We know by
induction hypothesis that Rk ⊆ R, which means that (c, b) ∈ R. Since R is transitive, and
(a, c) ∈ R and (c, b) ∈ R, we have (a, b) ∈ R. This completes the proof.

Mathematical Foundations of Computer Science Lecture Outline

Uploaded by

Copyright:

Available Formats

Mathematical Foundations of Computer Science Lecture Outline

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Mathematical Foundations of Computer Science Lecture Outline

Uploaded by

Copyright:

Available Formats

Mathematical Foundations of Computer Science

The Probabilistic Method

E[X ∪ Y ] = E[X] + E[Y ] (1)

We consider the following random variables.

Plugging the values of E[X] and E[Y ] in (1) we get

E[X ∪ Y ] ≤ np + n(1 − p)δ+1 ≤ np + ne−p(δ+1) ,

The last expression is minimized when

Example. Prove that if n is large enough then there is a k-dominated tournament on n

Example. A relation R on a set A is symmetric iff R = R−1 .

Solution. ( =⇒ ) Suppose R is symmetric on A. We will prove that R = R−1 by showing

S ◦ R = {(x, z) | there exists a y ∈ B such that x R y and y S z}

What is the composite of the relations R and S?

Let R be a relation on a set A. The powers Rn , n = 1, 2, 3, . . . , are defined recursively by

Observe that R2 = R ◦ R, R3 = R2 ◦ R = (R ◦ R) ◦ R, and so on.

Example. Let R be a relation on a set A. Then R is transitive iff Rn ⊆ R, for all n ≥ 1.

You might also like