Discrete Mathematics
COMS 3203
Zeph Grunschlag
Copyright © Zeph Grunschlag,
2001-2002.
Agenda
Course policies
Quick Overview
Logic
Lecture 1
2
About me
3rd year at Columbia
Graduated from Berkeley in ’99


Mathematics PhD.
Thesis title: Algorithms in Geometric Group Theory
Research areas:



Lecture 1
Group theory and topology
Formal languages
Computer vision
3
Course Policies, Schedule,
Syllabus
Handout explaining required text,
overview of course, grading breakdown,
policies, etc. –The “contract”
Available from course homepage:
http://www.cs.columbia.edu/~zeph/3203
All homework, and most other lecture
material will be available online
Lecture 1
4
Lectures and Notes
Lectures are electronic and will be
available online after class.
There will usually be a low-tech
component (blackboard + chalk) which
you are responsible for.
Lecture 1
5
Quick Overview
Discrete Math is essentially that branch of
mathematics that does not depend on
limits; in this sense, it is the antithesis of Calculus. As computers are
discrete object operating one jumpy,
discontinuous step at a time, Discrete
Math is the right framework for
describing precisely Computer Science
concepts.
Lecture 1
6
Quick Overview
The conceptual center of computer
science is the ALGORITHM.
Lecture 1
7
Quick Overview
Discrete Math helps provide…
…the
machinery necessary
for creating sophisticated
algorithms
…the
tools for analyzing their
efficiency
…the means of proving their validity
Lecture 1
8
Quick Overview
Although the point of 3203 is to provide
the tools for creating and analyzing
sophisticated algorithms, we won’t
focus on the algorithmic aspect, with
some exceptions. To see the
algorithmic applications take almost any
4000-level CS course or 3137/9 (Data
Structures).
Lecture 1
9
Quick Overview - Topics
Logic and Sets



Make notions you’re already used to from
programming a little more rigorous (operators)
Fundamental to all mathematical disciplines
Useful for digital circuits, hardware design
Elementary Number Theory


Lecture 1
Get to rediscover the old reliable number and
find out some surprising facts
Very useful in crypto-systems
10
Quick Overview - Topics
Proofs (especially induction)


If you want to debug a program beyond a
doubt, prove that it’s bug-free
Proof-theory has recently also been shown to be
useful in discovering bugs in pre-production
hardware
Counting and Combinatorics



Lecture 1
Compute your odds of winning lottery
Important for predicting how long certain
computer program will take to finish
Useful in designing algorithms
11
Quick Overview - Topics
Graph Theory


Many clever data-structures for organizing
information and making programs highly
efficient are based on graph theory
Very useful in describing problems in




Lecture 1
Databases
Operating Systems
Networks
EVERY CS DISCIPLINE!!!!
12
Section 1.1: Logic
Axiomatic concepts in math:
Equals
Opposite
Truth and falsehood
Statement
Objects
Collections
Lecture 1
13
Section 1.1: Logic
We intuitively know that Truth and Falsehood
are opposites. That statements describe the
world and can be true/false. That the world
is made up of objects and that objects can be
organized to form collections.
The foundations of logic mimic our intuitions by
setting down constructs that behave
analogously.
Lecture 1
14
False, True, Statements
Axiom: False is the opposite to Truth.
A statement is a description of something.
Examples of statements:




I’m 31 years old.
I have 17 children.
I always tell the truth.
I’m lying to you.
Q’s: Which statements are True? False? Both?
Neither?
Lecture 1
15
False, True, Statements
True: I’m 31 years old.
False: I have 17 children.
I always tell the truth.
Both: IMPOSSIBLE, by our Axiom.
Lecture 1
16
False, True, Statements
Neither: I’m lying to you. (If viewed on its own)
HUH? Well suppose that
S = “I’m lying to you.”
were true. In particular, I am actually lying, so
S is false. So it’s both true and false,
impossible by the Axiom.
Okay, so I guess S must be false. But then I
must not be lying to you. So the statement is
true. Again it’s both true and false.
In both cases we get the opposite of our
assumption, so S is neither true nor false.
Lecture 1
17
Propositions
To avoid painful head-aches, we ban such
silly non-sense and avoid the most
general type of statements limiting
ourselves to statements with valid
truth-values instead:
DEF: A proposition is a statement that is
true or false.
Lecture 1
18
Propositions
Propositional Logic is a static discipline of
statements which lack semantic content.
E.G. p = “Clinton was the president.”
q = “The list of U.S. presidents includes
Clinton.”
r = “Lions like to sleep.”
All p and q are no more closely related than q
and r are, in propositional calculus. They are
both equally related as all three statements
are true. Semantically, however, p and q are
the same!
Lecture 1
19
Propositions
So why waste time on such matters?
Propositional logic is the study of how simple
propositions can come together to make
more complicated propositions. If the simple
propositions were endowed with some
meaning –and they will be very soon– then
the complicated proposition would have
meaning as well, and then finding out the
truth value is actually important!
Lecture 1
20
Compound Propositions
In Propositional Logic, we assume a
collection of atomic propositions are
given: p, q, r, s, t, ….
Then we form compound propositions by
using logical connectives (logical
operators) to form propositional
“molecules”.
Lecture 1
21
Logical Connectives
Operator
Negation
Conjunction
Disjunction
Exclusive or
Conditional
Biconditional
Lecture 1
Symbol Usage






Java
not
!
and
&&
or
||
xor
(p||q)&&(!p||!q)
if,then
p?q:true
iff
(p&&q)||(!p&&!q)
22
Compound Propositions:
Examples
p = “Cruise ships only go on big rivers.”
q = “Cruise ships go on the Hudson.”
r = “The Hudson is a big river.”
r = “The Hudson is not a big river.”
pq = “Cruise ships only go on big rivers and
go on the Hudson.”
pq r = “If cruise ships only go on big rivers
and go on the Hudson, then the Hudson is a
big river.”
Lecture 1
23
Negation
This just turns a false proposition to true
and the opposite for a true proposition.
EG: p = “23 = 15 +7”
p happens to be false, so p is true.
In Java, “!” plays the same role:
!(23 == 15+7)
has the boolean value true whenever
evaluated.
Lecture 1
24
Negation – truth table
Logical operators are defined by truth
tables –tables which give the output of
the operator in the right-most column.
Here is the truth table for negation:
p
F
T
Lecture 1
p
T
F
25
Conjunction
Conjunction is a binary operator in that it
operates on two propositions when
creating compound proposition. On the
other hand, negation is a unary
operator (the only non-trivial one
possible).
Lecture 1
26
Conjunction
Conjunction is supposed to encapsulate
what happens when we use the word
“and” in English. I.e., for “p and q ” to
be true, it must be the case that BOTH
p is true, as well as q. If one of these
is false, than the compound statement
is false as well.
Lecture 1
27
Conjunction
EG. p = “Clinton was the president.”
q = “Monica was the president.”
r = “The meaning of is is important.”
Assuming p and r are true, while q false.
Out of pq, pr, qr
only pr is true.
Java: x==3 && x!=3
Evaluates to false for any possible value of x.
Lecture 1
28
Conjunction – truth table
Lecture 1
p
q
p q
T
T
F
F
T
F
T
F
T
F
F
F
29
Disjunction – truth table
Conversely, disjunction is true when at
least one of the components is true:
Lecture 1
p
q
p q
T
T
F
F
T
F
T
F
T
T
T
F
30
Disjunction – caveat
Note: English version of disjunction “or”
does not always satisfy the assumption
that one of p/q being true implies that
“p or q ” is true.
Q: Can someone come up with an
example?
Lecture 1
31
Disjunction – caveat
A: The entrée is served with
soup or salad.
Most restaurants definitely don’t allow
you to get both soup and salad so that
the statement is false when both soup
and salad is served. To address this
situation, exclusive-or is introduced
next.
Lecture 1
32
Exclusive-Or – truth table
p
q
p q
T
T
F
F
T
F
T
F
F
T
T
F
Note: in this course any usage of “or”
will connote the logical operator 
as opposed to the exclusive-or.
Lecture 1
33
Conditional (Implication)
This one is probably the least intuitive.
It’s only partly akin to the English usage
of “if,then” or “implies”.
DEF: p  q is true if q is true, or if p is
false. In the final case (p is true while
q is false) p  q is false.
Semantics: “p implies q ” is true if one
can mathematically derive q from p.
Lecture 1
34
Conditional -- truth table
Lecture 1
p
q
p q
T
T
F
F
T
F
T
F
T
F
T
T
35
Conditional
Q: Does this makes sense? Let’s try
examples for each row of truth table:
1. If pigs like mud then pigs like mud.
2. If pigs like mud then pigs can fly.
3. If pigs can fly then pigs like mud.
4. If pigs can fly then pigs can fly.
Lecture 1
36
Conditional
A:
1. If pigs like mud then pigs like mud.
True: nothing about this statement is false.
2. If pigs like mud then pigs can fly.
False: seems to assert falsehood
3. If pigs can fly then pigs like mud.
True: argument for –only care about end-result.
Argument against –counters common English
hyperbole.
4. If pigs can fly then pigs can fly.
True. WAIT! By first reasoning in 3, when “if” part is
false, should only care about “then” part!!!!!
On other hand, standard English hyperbole.
Lecture 1
37
Conditional: why FF is True
Remember, all of these are mathematical
constructs, not attempts to mimic English.
Mathematically, p should imply q whenever it
is possible to derive q by from p by using
valid arguments. For example consider the
mathematical analog of no. 4:
If 0 = 1 then 3 = 9.
Q: Is this true mathematically?
Lecture 1
38
Conditional: why FF is True
A: YES mathematically and YES by the
truth table.
Here’s a mathematical proof:
1. 0 = 1 (assumption)
Lecture 1
39
Conditional: why FF is True
A: YES mathematically and YES by the
truth table.
Here’s a mathematical proof:
1. 0 = 1 (assumption)
2. 1 = 2 (added 1 to both sides)
Lecture 1
40
Conditional: why FF is True
A: YES mathematically and YES by the
truth table.
Here’s a mathematical proof:
1. 0 = 1 (assumption)
2. 1 = 2 (added 1 to both sides)
3. 3 = 6 (multiplied both sides by 3)
Lecture 1
41
Conditional: why FF is True
A: YES mathematically and YES by the
truth table.
Here’s a mathematical proof:
1. 0 = 1 (assumption)
2. 1 = 2 (added 1 to both sides)
3. 3 = 6 (multiplied both sides by 3)
4. 0 = 3 (multiplied no. 1 by 3)
Lecture 1
42
Conditional: why FF is True
A: YES mathematically and YES by the
truth table.
Here’s a mathematical proof:
1. 0 = 1 (assumption)
2. 1 = 2 (added 1 to both sides)
3. 3 = 6 (multiplied both sides by 3)
4. 0 = 3 (multiplied no. 1 by 3)
5. 3 = 9 (added no. 3 and no. 4)
Lecture 1
QED
43
Conditional: why FF is True
As we want the conditional to make sense
in the semantic context of
mathematics, we better define it as
we have!
Other questionable rows of the truth table
can also be justified in a similar
manner.
Lecture 1
44
Conditional: synonyms
There are many ways to express the conditional
statement p  q :
If p then q. p implies q. If p, q.
p only if q. p is sufficient for q.
Some of the ways reverse the order of p and q
but have the same connotation:
q if p. q whenever p. q is necessary for p.
To aid in remembering these, I suggest
inserting “is true” after every variable:
EG: “p is true only if q is true”
Lecture 1
45
Bi-Conditional -- truth table
For p  q to be true, p and q must have
the same truth value. Else, p
p
q
pq
T
T
F
F
T
F
T
F
T
F
F
T
Q : Which operator is
Lecture 1
 q is false:
 the opposite of?
46
Bi-Conditional
A:
 has exactly the opposite truth table
as .
This means that we could have defined the
bi-conditional in terms of other previously
defined symbols, so it is redundant. In fact,
only really need negation and disjunction to
define everything else.
Extra operators are for convenience.
Q: Could we define all other logical operations
using only negation and exclusive or?
Lecture 1
47
Bi-Conditional
A: No. Notice that negation and
exclusive-or each maintain parity
between truth and false: No matter
what combination of these symbols,
impossible to get a truth table with four
output rows consisting of 3 T’s and 1 F
(such as implication and disjuction).
Lecture 1
48
Bit Strings
Electronic computers achieve their calculations
inside semiconducting materials. For
reliability, only two stable voltage states are
used and so the most fundamental operations
are carried out by switching voltages between
these two stable states.
In logic, only two truth values are allowed.
Thus propositional logic is ideal for modeling
computers. High voltage values are modeled
by True, which for brevity we call the number
1, while low voltage values are modeled by
False or 0.
Lecture 1
49
Bit Strings
Thus voltage memory stored in a computer can
be represented by a sequence of 0’s and 1’s
such as
01 1011 0010 1001
Another portion of the memory might look like
10 0010 1111 1001
Each of the number in the sequence is called a
bit, and the whole sequence of bits is called
a bit string.
Lecture 1
50
Bit Strings
It turns out that the analogs of the logical
operations can be carried out quite easily
inside the computer, one bit at a time. This
can then be transferred to whole bit strings.
For example, the exclusive-or of the previous
bit strings is:
01 1011 0010 1001

Lecture 1
10 0010 1111 1001
11 1001 1101 0000
51
Blackboard Exercises for 1.1
Worked out on the black-board.
1.1.6.c, 1.1.22.d, 1.1.39:
1. q = “You miss the final exam.”
r = “You pass the course.”
Express q  r in English.
2. Construct a truth table for p q.
3. Can one determine relative salaries of F
(Fannie), J (Janice) and M (Maggie) from
the following?
1.
2.
Lecture 1
If F is not highest paid, then J is.
If J is not lowest paid, then M is highest paid.
52
Descargar

Machine Models - Columbia University