TECH

Computer Science
CS520 Advanced Analysis of Algorithms and
Complexity
Dr. Ben Choi
Ph.D. in EE (Computer Engineering),
The Ohio State University
System Performance Engineer,
Lucent Technologies - Bell Labs Innovations
Pilot, FAA certified pilot for
airplanes and helicopters

BenChoi.info

What is a Computer Algorithm?

A computer algorithm is
a detailed step-by-step method for
solving a problem
by using a computer.

Problem-Solving (Science and Engineering)
Analysis
How does it work?
Breaking a system down to known components
How the components relate to each other
Breaking a process down to known functions

Synthesis
Building tools and toys!
What components are needed
How the components should be put together
Composing functions to form a process

Problem Solving Using Computers
Problem:
Strategy:
Algorithm:
Input:
Output:
Step:
Analysis:
Correctness:
Time & Space:
Optimality:
Implementation:
Verification:
Example: Search in an unordered array
Problem:
Let E be an array containing n entries, E[0], , E[n-1],
in no particular order.
Find an index of a specified key K, if K is in the array;
return 1 as the answer if K is not in the array.
Strategy:
Compare K to each entry in turn until a match is found
or the array is exhausted.
If K is not in the array, the algorithm returns 1 as its
answer.
Example: Sequential Search, Unordered
Algorithm (and data structure)
Input: E, n, K, where E is an array with n entries
(indexed 0, , n-1), and K is the item sought. For
simplicity, we assume that K and the entries of E are
integers, as is n.
Output: Returns ans, the location of K in E
(-1 if K is not found.)

Algorithm: Step (Specification)
int seqSearch(int[] E, int n, int K)
1. int ans, index;
2. ans = -1; // Assume failure.
3. for (index = 0; index < n; index++)
4. if (K == E[index])
5. ans = index; // Success!
6. break; // Done!
7. return ans;
Analysis of the Algorithm
How shall we measure the amount of work done by
an algorithm?
Basic Operation:
Comparison of x with an array entry
Worst-Case Analysis:
Let W(n) be a function. W(n) is the maximum number
of basic operations performed by the algorithm on any
input size n.
For our example, clearly W(n) = n.
The worst cases occur when K appears only in the last
position in the array and when K is not in the array at
all.
Average-Behavior Analysis:
Let q be the probability that K is in the array
A(n) = n(1 q) + q
Optimality:
The Best possible solution?
Searching an Ordered Array
Using Binary Search
W(n) = Ceiling[lg(n+1)]
The Binary Search algorithm is optimal.
Correctness: (Proving Correctness of Procedures s3.5)
More Analysis of the Algorithm
( )
(
1 lg + = n
What is CS 520?
Class Syllabus
Algorithm Language (Specifying the Steps)
Java as an algorithm language
Syntax similar to C++
Some steps within an algorithm may be specified in
pseduocode (English phrases)
Focus on the strategy and techniques of an algorithm,
not on detail implementation
Analysis Tool: Mathematics: Set
A set is a collection of distinct elements.
The elements are of the same type, common
properties.
element e is a member of set S is denoted as
e e S
Read e is in S
A particular set is defined by listing or describing its
elements between a pair of curly braces:
S
1
= {a, b, c}, S
2
= {x | x is an integer power of 2}
read the set of all elements x such that x is
S
3
= {} = C, has not elements, called empty set
A set has no inherent order.
Subset, Superset; Intersection, Union
If all elements of one set, S
1
are also in another set, S
2
,
Then S
1
is said to be a subset of S
2
, S
1
_ S
2
and S
2
is said to be a superset of S
1
, S
2
_ S
1
.
Empty set is a subset of every set, C _ S
Intersection
S T = {x | x e S and x e T}
Union
S T = {x | x e S or x e T}
Cardinality
Cardinality
A set, S, is finite if there is an integer n such that the
elements of S can be placed in a one-to-one
correspondence with {1, 2, 3, , n}
in this case we write |S| = n
How many distinct subsets does a finite set on n
elements have? There are 2
n
subsets.
How many distinct subsets of cardinality k does a
finite set of n elements have?
There are C(n, k) = n!/((n-k)!k!), n choose k
|
|
.
|

\
|
k
n
Sequence
A group of elements in a specified order is called a
sequence.
A sequence can have repeated elements.
Sequences are defined by listing or describing their
elements in order, enclosed in parentheses.
e.g. S1 = (a, b, c), S2 = (b, c, a), S3 = (a, a, b, c)
A sequence is finite if there is an integer n such that
the elements of the sequence can be placed in a one-
to-one correspondence with (1, 2, 3, , n).
If all the elements of a finite sequence are distinct,
that sequence is said to be a permutation of the finite
set consisting of the same elements.
One set of n elements has n! distinct permutations.
Tuples and Cross Product
A tuple is a finite sequence.
Ordered pair (x, y), triple (x, y, z),
quadruple, and quintuple
A k-tuple is a tuple of k elements.
The cross product of two sets, say S and T, is
S T = {(x, y) | x e S, y e T}
| S T | = |S| |T|
It often happens that S and T are the same set, e.g.
N N
where N denotes the set of natural numbers,
{0,1,2,}
Relations and Functions
A relation is some subset of a (possibly iterated) cross product.
A binary relation is some subset of a cross product,
e.g. R _ S T
e.g. less than relation can be defined as
{(x, y) | x e N, y e N, x < y}
Important properties of relations; let R _ S S
reflexive: for all x e S, (x, x) e R.
symmetric: if (x, y) e R, then (y, x) e R.
antisymmetric: if (x, y) e R, then (y, x) e R
transitive: if (x,y) e R and (y, z) e R, then (x, z) e R.
A relation that is reflexive, symmetric, and transitive is called
an equivalence relation, partition the underlying set S into
equivalence classes [x] = {y e S | x R y}, x e S
A function is a relation in which no element of S (of S x T) is
repeated with the relation. (informal def.)
Analysis Tool: Logic
Logic is a system for formalizing natural language
statements so that we can reason more accurately.
The simplest statements are called atomic formulas.
More complex statements can be build up through the
use of logical connectives: . and, v or, not,
implies A B A implies B if A then B
A B is logically equivalent to A v B
(A . B) is logically equivalent to A v B
(A v B) is logically equivalent to A . B
Quantifiers: all, some
for all x x P(x) is true iff P(x) is true for all x
universal quantifier (universe of discourse)
there exist x -x P(x) is true iff P(x) is true for some
value of x
existential quantifier
x A(x) is logically equivalent to -x(A(x))
-x A(x) is logically equivalent to x(A(x))
x (A(x) B(x))
For all x such that if A(x) holds then B(x) holds
Prove: by counterexample, Contraposition,
Contradiction
Counterexample
to prove x (A(x) B(x)) is false, we show some
object x for which A(x) is true and B(x) is false.
(x (A(x) B(x))) -x (A(x) . B(x))
Contraposition
to prove A B, we show ( B) ( A)
Contradiction
to prove A B, we assume B and then prove B.
A B (A . B) B
A B (A . B) is false
Assuming (A . B) is true,
and discover a contradiction (such as A . A),
then conclude (A . B) is false, and so A B.
Prove: by Contradiction, e.g.
Prove [B . (B C)] C
by contradiction
Proof:
Assume C
C . [B . (B C)]
C . [B . ( B v C)]
C . [(B . B) v (B . C)]
C . [(B . C)]
C . C . B
False, Contradiction
C
Rules of Inference
A rule of inference is a general pattern that allows us
to draw some new conclusion from a set of given
statements.
If we know {} then we can conclude {}
If {B and (B C)} then {C}
modus ponens
If {A B and B C} then {A C}
syllogism
If {B C and B C} then {C}
rule of cases
Two-valued Boolean (algebra) logic
1. There exists two elements in B, i.e. B={0,1}
there are two binary operations + or, v, and, .
2. Closure: if x, y e B and z = x + y then z e B
if x, y e B and z = x y then z e B
3. Identity element: for + designated by 0: x + 0 = x
for designated by 1: x 1 = x
4. Commutative: x + y = y + x
x y = y x
5. Distributive: x (y + z) = (x y) + (x z)
x + (y z) = (x + y) (x + z)
6. Complement: for every element x e B,
there exits an element x e B
x + x = 1, x x = 0
True Table and Tautologically Implies e.g.
Show [B . (B C)] C is a tautology:

B C (B C) [B . (B C)] [B . (B C)] C
0 0 1 0 1
0 1 1 0 1
1 0 0 0 1
1 1 1 1 1

For every assignment for B and C,
the statement is True
Prove: by Rule of inferences
Prove [B . (B C)] C
Proof:
[B . (B C)] C
[B . (B C)] v C
[B . ( B v C)] v C
[(B . B) v (B . C)] v C
[(B . C)] v C
B v C v C
True (tautology)
Direct Proof:
[B . (B C)] [B . C] C
Analysis Tool: Probability
Elementary events (outcomes)
Suppose that in a given situation an event, or
experiment, may have any one, and only one, of k
outcomes, s
1
, s
2
, , s
k
. (mutually exclusive)
Universe
The set of all elementary events is called the universe
and is denoted U = {s
1
, s
2
, , s
k
}.
Probability of s
i

associate a real number Pr(s
i
), such that
0 s Pr(s
i
) s 1 for 1 s i s k;
Pr(s
1
) + Pr(s
2
) + + Pr(s
k
) = 1
Event
Let S _ U. Then S is called an event, and
Pr(S) =
si e S
Pr(s
i
)
Sure event U = {s
1
, s
2
, , s
k
}, Pr(U) = 1
Impossible event, C ,Pr(C) = 0
Complement event not S U S,
Pr(not S) = 1 Pr(S)
Conditional Probability
The conditional probability of an event S given an
event T is defined as
Pr(S | T) = Pr(S and T) / Pr(T)
=
si e ST
Pr(s
i
) /
sj e T
Pr(s
j
)
Independent
Given two events S and T, if
Pr(S and T) = Pr(S)Pr(T)
then S and T are stochastically independent, or simply
independent.
Random variable and their Expected value
A random variable is a real valued variable that
depends on which elementary event has occurred
it is a function defined for elementary events.
e.g. f(e) = the number of inversions in the permutation
of {A, B, C}; assume all input permutations are equally
likely.
Expectation
Let f(e) be a random variable defined on a set of
elementary events e e U. The expectation of f, denoted
as E(f), is defined as
E(f) =
e e U
f(e)Pr(e)
This is often called the average values of f.
Expectations are often easier to manipulate then the
random variables themselves.
Conditional expectation and
Laws of expectations
The conditional expectation of f given an event S,
denoted as E (f | S), is defined as
E(f | S) =
e e S
f(e)Pr(e | S)
Law of expectations
For random variables f(e) and g(e) defined on a set of
elementary events e e U, and any event S:
E(f + g) = E(f) + E(g)
E(f) = Pr(S)E(f | S) + Pr(not S) E(f | not S)
Analysis Tool: Algebra
Manipulating Inequalities
Transitivity: If ((A s B) and (B s C) Then (A s C)
Addition: If ((A s B) and (C s D) Then (A+C s B+D)
Positive Scaling:
If ((A s B) and (o > 0) Then (o A s o B)
Floor and Ceiling Functions
Floor[x] is the largest integer less than or equal to x.

Ceiling[x] is the smallest integer greater than or equal
to x.

x
(
x
Logarithms
For b>1 and x>0,
log
b
x (read log to the base b of x)
is that real number L such that b
L
= x
log
b
x is the power to which b must be raised to get x.
Log properties: def: lg x = log
2
x; ln x = log
e
x
Let x and y be arbitrary positive real numbers, let a, b any real
number, and let b>1 and c>1 be real numbers.
log
b
is a strictly increasing function,
if x > y then log
b
x > log
b
y
log
b
is a one-to-one function,
if log
b
x = log
b
y then x = y
log
b
1 = 0; log
b
b = 1; log
b
x
a
= a log
b
x
log
b
(xy) = log
b
x + log
b
y
x
log y
= y
log x

change base:
log
c
x = (log
b
x)/(log
b
c)
Series
A series is the sum of a sequence.
Arithmetic series
The sum of consecutive integers
Polynomial Series
The sum of squares
The general case is

Power of 2

Arithmetic-
Geometric Series
2
) 1 (
1
+
=

=
n n
i
n
i
3 6
3 2
3 2 3
1
2
n n n n
i
n
i
~
+ +
=

=
1
1
1
+
~
+
=

k
n
i
k
n
i
k
1 2 2
1
0
=
+
=

k
k
i
i
( ) 2 2 1 2
1
1
+ =
+
=

k
k
i
i
k i
Summations Using Integration
A function f(x) is said to be monotonic, or
nondecreasing, if x s y always implies that f(x) s f(y).
A function f(x) is antimonotonic, or nonincreasing,
if f(x) is monotonic.
If f(x) is nondecreasing then



If f(x) is nonincreasing then

}

}
+
=

s s
1
1
) ( ) ( ) (
b
a
b
a i
b
a
dx x f i f dx x f
}

}

=
+
s s
b
a
b
a i
b
a
dx x f i f dx x f
1
1
) ( ) ( ) (
Classifying functions by their
Asymptotic Growth Rates
asymptotic growth rate, asymptotic order, or
order of functions
Comparing and classifying functions that ignores
constant factors and small inputs.
The Sets big oh O(g), big theta O(g), big omega O(g)
g
O(g): functions that grow at least as fast as g
O(g): functions that grow at the same rate as g
O(g): functions that grow no faster than g
The Sets O(g), O(g), O(g)
Let g and f be a functions from
the nonnegative integers into the positive real numbers
For some real constant c > 0 and
some nonnegative integer constant n
0
O(g) is the set of functions f, such that
f(n) s c g(n) for all n > n
0
O(g) is the set of functions f, such that
f(n) > c g(n) for all n > n
0
O(g) = O(g) O(g)
asymptotic order of g
f eO(g) read as
f is asymptotic order g or f is order g
Comparing asymptotic growth rates
Comparing f(n) and g(n) as n approaches infinity,
IF


< , including the case in which the limit is 0 then
f eO(g)
> 0, including the case in which the limit is then
f e O(g)
= c and 0 < c < then
f e O(g)
= 0 then f e o(g) //read as little oh of g
= then f e e(g) //read as little omega of g
) (
) (
lim
n g
n f
n
Properties of O(g), O(g), O(g)
Transitive: If f eO(g) and g eO(h), then f eO(h)
O is transitive. Also O, O, o, e are transitive.
Reflexive: f e O(f)
Symmetric: If f e O(g), then g e O(f)
O defines an equivalence relation on the functions.
Each set O(f) is an equivalence class (complexity class).
f eO(g) g e O(f)
O(f + g) = O(max(f, g))
similar equations hold for O and O
Classification of functions, e.g.
O(1) denotes the set of functions bounded by a
constant (for large n)
f e O(n), f is linear
f e O(n
2
), f is quadratic; f e O(n
3
), f is cubic
lg n e o(n
o
) for any o > 0, including factional powers
n
k
e o(c
n
) for any k > 0 and any c > 1
powers of n grow more slowly than
any exponential function c
n
) (
1
1
+
=
O e

d
n
i
d
n i
n of function some be may b 1, r 0, r for ) ( = > O e

=
b
b
a i
i
r r
)) log( ( ) log(
1
n n i
n
i
O e

=
Analyzing Algorithms and Problems
We analyze algorithms with the intention of
improving them, if possible, and
for choosing among several available for a problem.

Correctness
Amount of work done, and space used
Optimality, Simplicity
Correctness can be proved!
An algorithm consists of sequences of steps
(operations, instructions, statements) for transforming
inputs (preconditions) to outputs (postconditions)
Proving
if the preconditions are satisfied,
then the postconditions will be true,
when the algorithm terminates.
Amount of work done
We want a measure of work that tells us something
about the efficiency of the method used by the algorithm
independent of computer, programming language,
programmer, and other implementation details.
Usually depending on the size of the input
Counting passes through loops
Basic Operation
Identify a particular operation fundamental to the
problem
the total number of operations performed is roughly
proportional to the number of basic operations
Identifying the properties of the inputs that affect the
behavior of the algorithm
Worst-case complexity
Let D
n
be the set of inputs of size n for the problem
under consideration, and let I be an element of D
n
.
Let t(I) be the number of basic operations performed by
the algorithm on input I.
We define the function W by
W(n) = max{t(I) | I e D
n
}
called the worst-case complexity of the algorithm.
W(n) is the maximum number of basic operations
performed by the algorithm on any input of size n.
The input, I, for which an algorithm behaves worst
depends on the particular algorithm.
Average Complexity
Let Pr(I) be the probability that input I occurs.
Then the average behavior of the algorithm is defined as
A(n) = E
I e Dn
Pr(I) t(I).
We determine t(I) by analyzing the algorithm,
but Pr(I) cannot be computed analytically.
A(n) = Pr(succ)A
succ
(n) + Pr(fail)A
fail
(n)
An element I in D
n
may be thought as a set or
equivalence class that affect the behavior of the
algorithm. (see following e.g. n+1 cases)
e.g. Search in an unordered array
int seqSearch(int[] E, int n, int K)
1. int ans, index;
2. ans = -1; // Assume failure.
3. for (index = 0; index < n; index++)
4. if (K == E[index])
5. ans = index; // Success!
6. break; // Done!
7. return ans;
Average-Behavior Analysis e.g.
A(n) = Pr(succ)A
succ
(n) + Pr(fail)A
fail
(n)
There are total of n+1 cases of I in D
n
Let K is in the array as succ cases that have n cases.
Assuming K is equally likely found in any of the n
location, i.e. Pr(I
i
| succ) = 1/n
for 0 <= i < n, t(I
i
) = i + 1
A
succ
(n) = E
i=0
n-1
Pr(I
i
| succ) t(I
i
)
= E
i=0
n-1
(1/n) (i+1) = (1/n)[n(n+1)/2] = (n+1)/2
Let K is not in the array as the fail case that has 1
cases, Pr(I | fail) = 1
Then A
fail
(n) = Pr(I | fail) t(I) = 1 n
Let q be the probability for the succ cases
q [(n+1)/2] + (1-q) n
Space Usage
If memory cells used by the algorithms depends on
the particular input,
then worst-case and average-case analysis can be done.
Time and Space Tradeoff.
Optimality the best possible
Each problem has inherent complexity
There is some minimum amount of work required to
solve it.
To analyze the complexity of a problem,
we choose a class of algorithms, based on which
prove theorems that establish a lower bound on the
number of operations needed to solve the problem.
Lower bound (for the worst case)
Show whether an algorithm is optimal?
Analyze the algorithm, call it A, and found the Worst-
case complexity W
A
(n), for input of size n.
Prove a theorem starting that,
for any algorithm in the same class of A
for any input of size n, there is some input for which the
algorithm must perform
at least W
[A]
(n)
(lower bound in the worst-case)
If W
A
(n) == W
[A]
(n)
then the algorithm A is optimal
else there may be a better algorithm
OR there may be a better lower bound.
Optimality e.g.
Problem
Fining the largest entry in an (unsorted) array of
n numbers
Algorithm A
int findMax(int[] E, int n)
1. int max;
2. max = E[0]; // Assume the first is max.
3. for (index = 1; index < n; index++)
4. if (max < E[index])
5. max = E[index];
6. return max;
Analyze the algorithm, find W
A
(n)
Basic Operation
Comparison of an array entry with another array entry
or a stored variable.
Worst-Case Analysis
For any input of size n, there are exactly n-1 basic
operations
W
A
(n) = n-1
For the class of algorithm [A], find W
[A]
(n)
Class of Algorithms
Algorithms that can compare and copy the numbers,
but do no other operations on them.
Finding (or proving) W
[A]
(n)
Assuming the entries in the array are all distinct
(permissible for finding lower bound on the worst-case)
In an array with n distinct entries, n 1 entries are not
the maximum.
To conclude that an entry is not the maximum, it must
be smaller than at least one other entry. And, one
comparison (basic operation) is needed for that.
So at least n-1 basic operations must be done.
W
[A]
(n) = n 1
Since W
A
(n) == W
[A]
(n), algorithm A is optimal.
Simplicity
Simplicity in an algorithm is a virtue.
Designing Algorithms
Problem solving using Computer
Algorithm Design Techniques
divide-and-conquer
greedy methods
depth-first search (for graphs)
dynamic programming
Problem and Strategy A
Problem: array search
Given an array E containing n and given a value K, find
an index for which K = E[index] or, if K is not in the
array, return 1 as the answer.
Strategy A
Input data and Data structure: unsorted array
sequential search
Algorithm A
int seqSearch(int[] E, int n, int k)
Analysis A
W(n) = n
A(n) = q [(n+1)/2] + (1-q) n
Better Algorithm and/or Better Input Data
Optimality A
for searching an unsorted array
W
[A]
(n) = n
Algorithm A is optimal.
Strategy B
Input data and Data structure: array sorted in
nondecreasing order
sequential search
Algorithm B.
int seqSearch(int[] E, int n, int k)
Analysis B
W(n) = n
A(n) = q [(n+1)/2] + (1-q) n
Better Algorithm
Optimality B
It makes no use of the fact that the entries are ordered
Can we modify the algorithm so that it uses the added
information and does less work?
Strategy C
Input data and Data structure: array sorted in
nondecreasing order
sequential search:
as soon as an entry larger than K is encountered, the
algorithm can terminate with the answer 1.
Algorithm C: modified sequential search
int seqSearchMod(int[] E, int n, int K)
1. int ans, index;
2. ans = -1; // Assume failure.
3. for (index = 0; index < n; index++)
4. if (K > E[index])
5. continue;
6. if (K < E[index])
7. break; // Done!
8. // K == E[index]
9. ans = index; // Find it
10. break;
11. return ans;
Analysis C
W(n) = n + 1 ~ n
Average-Behavior
n cases for success:
A
succ
(n) = E
i=0
n-1
Pr(I
i
| succ) t(I
i
)
= E
i=0
n-1
(1/n) (i+2) = (3 + n)/2
n+1 cases or (gaps) for fail: <E[0]<>E[1]E[n-1]>
A
fail
(n) = Pr(I
i
| fail) t(I
i
) =
E
i=0
n-1
(1/(n+1)) (i+2) + n/(n+1)
A(n) = q (3+n)/2 + (1-q) ( n /(n+1) + (3+n)/2 )
~ n/2
Lets Try Again! Lets divide-and-conquer!
Strategy D
compare K first to the entry in the middle of the array
-- eliminates half of the entry with one comparison
apply the same strategy recursively
Algorithm D: Binary Search
Input: E, first, last, and K, all integers, where E is an
ordered array in the range first, , last, and K is the
key sought.
Output: index such that E[index] = K if K is in E within
the range first, , last, and index = -1 if K is not in this
range of E
Binary Search
int binarySearch(int[] E, int first, int last, int K)
1. if (last < first)
2. index = -1;
3. else
4. int mid = (first + last)/2
5. if (K == E[mid])
6. index = mid;
7. else if (K < E[mid])
8. index = binarySearch(E, first, mid-1, K)
9. else
10. index = binarySearch(E, mid+1, last, K);
11. return index
Worst-Case Analysis of Binary Search
Let the problem size be n = last first + 1; n>0
Basic operation is a comparison of K to an array entry
Assume one comparison is done with the three-way
branch
First comparison, assume K != E[mid], divides the
array into two sections, each section has at most
Floor[n/2] entries.
estimate that the size of the range is divided by 2 with
each recursive call
How many times can we divide n by 2 without getting a
result lest than 1 (i.e. n/(2
d
) >= 1) ?
d <= lg(n), therefore we do Floor[lg(n)] comparison
following recursive calls, and one before that.
W(n) = Floor[lg(n)] + 1 = Ceiling[lg(n + 1)] e O(log n)
Average-Behavior Analysis of Binary Search
There are n+1 cases, n for success and 1 for fail
Similar to worst-case analysis, Let n = 2
d
1
A
fail
= lg(n+1)
Assuming Pr(I
i
| succ) = 1/n for 1 <= i <= n
divide the n entry into groups, S
t
for 1 <= t <= d,
such that S
t
requires t comparisons
(capital S for group, small s for cardinality of S)
It is easy to see (?) that (members contained in the group)
s
1
= 1 = 2
0
, s
2
= 2 = 2
1
, s
3
= 4 = 2
2
, and
in general, s
t
= 2
t-1
The probability that the algorithm does t comparisons is s
t
/n
A
succ
(n) = E
t=1
d
(s
t
/n) t = ((d 1)2
d
+ 1)/n
d = lg(n+1)
A
succ
(n) = lg(n+1) 1 + lg(n+1)/n
A(n) ~ lg(n+1) q, where q is probability of successful search
Optimality of Binary Search
So far we improve from u(n) algorithm to u(log n)
Can more improvements be possible?
Class of algorithm: comparison as the basic operation
Analysis by using decision tree, that
for a given input size n is a binary tree whose nodes are
labeled with numbers between 0 and n-1 as e.g.
distance 0
distance 1
distance 2
distance 3
Decision tree for analysis
The number of comparisons performed in the worst
case is the number of nodes on a longest path from
the root to a leaf; call this number p.
Suppose the decision tree has N nodes
N <= 1 + 2 + 4 + + 2
p-1
N <= 2
p
1
2
p
>= (N + 1)
Claim N >= n if an algorithm A works correctly in all
cases
there is some node in the decision tree labeled i for each
i from 0 through n - 1
Prove by contradiction that N >= n
Suppose there is no node labeled i for some i in the
range from 0 through n-1
Make up two input arrays E1 and E2 such that
E1[i] = K but E2[i] = K > K
For all j < i, make E1[j] = E2[j] using some key values
less than K
For all j > i, make E1[j] = E2[j] using some key values
greater than K in sorted order
Since no node in the decision tree is labeled i, the
algorithm A never compares K to E1[i] or E2[i],
but it gives same output for both
Such algorithm A gives wrong output for at least one of
the array and it is not a correct algorithm
Conclude that the decision has at least n nodes
Optimality result
2
p
>= (N+1) >= (n+1)
p >= lg(n+1)

Theorem: Any algorithm to find K in an array of n
entries (by comparing K to array entries) must do at
least Ceiling[lg(n+1)] comparisons for some input.

Corollary: Since Algorithm D does Ceiling[lg(n+1)]
comparisons in the worst case, it is optimal.