Chapter 1

0 Notation
-

Let f and be functions If 7

g a
.

constant cent and a number K s -

t

flnlfcgcn ) b- nzk then we write

flnl =
Olgin) ) OR f =
0cg)

d- Notation
Let f and be functions If 7 constant
g
a
.

cent and a number KEN St flnlzegln)

n7k then we write
:
H
f- Mlg)
fln) =D lgln) ) OR

G- Notation
Let fond be functions if 7 two constants
g
.

s.tqglnkflnkczg.tn/-VnZk
G. GEIR and KEN
then we write :

fln / = ① ( gcn) ) OR F- ② (g)

Chapter 2

Sort Strassen Algo

Merge
Problem Sort
Problem Matrix Multiplication
:
:
array Input : Matrix M , and Mz of
Input :
Array of size n size nxn

Output : Sorted array output M , ✗ Mz:

Runtime :O / n
login) Runtime Olnlostd :
)
S, Biz Bzz SEA ,tA,z
-
=
,

2 I 4 3
53 Az, 1- Azz 54=132,1-13
-

2 1 9 13 •

2
I 4 3

11 2 314
Master theorem
.

1121314
'
Tlnl={a.TL%)
n -4
+ nd n >1

a > I b> I 020

Selection
Problem Find
Algosmallest d>
logbla) TIM -_
olndl
: kth
D= log , /a)
Input Arrays :
of size Tlnkolndlogln))
dclogbla)
n
and int k TlnI=O(n' 09 blay
, oaken
Output -
: K th -

smallest
elem ins
Runtime: Oln)

Finding Runtime :

Unfolding Induction :
Let be
Tlnl -11%1
n
large and assume
n Tlmlsc.vn for
+
all
k-mcn-tl.IT/+Z&n n

:
•
Chapter 3

Order
Depth-first Search / DFS) Topo
undirected
Problem : Explore directed
Problem : Explore graph
graphs
with vettices
Input Directed
: graph G
Graph C-
Input : a
w/ no
cycles
order of C-
✓ and
edges E Output Topo ;

Output None : A > C

n
F

Runtime: 0114+11=-1) r
n do B > D VE
A
d
89C
2
B) >
B→ A S
→,c→E
☐
3 14s 6 1 2
,
3
D E 4 5 g

DFS -

Forest compute Topo

116
1 : Run DFS
"
A
12%1 '
A 1215 Legend 2 : Run bucket sort
✓

? I C
¢ : tree edges on Postnumbers
/ q
✗ v10 \ e
'
: forward edges 3 : Reverse
posterder
,zyy T : back edges
4¥ Runtime :O / IVHIEI)
' F-
g. Cross edges
☐
l '

iii. ←

! -
'

sg6⇐ compute if graph

is cyclic
C- has directed A Run Dfslwlprefpostl
cycle
:
a
if
⇐> 2: _V non -
tree edge
Prelclcprecvkpostcvk Postal
DFS forest has
-
a
baekedge
Runtime : OWNED
 (

Chapter 3

Dijkstra's Algo
Problem Find : the shortest path between 2 vertices
in weighted directed
a
,
graph
1- input Directed
:

graph G w/ vertices V
and
edges Fe
,
where each
edge ahas a weight .

A source Vector SEV

Output :

the shortest path between s and vector VEV

any
Runtime ONE / +
XD logllvll)
:

Ys
at >✗
^
t ✗ y z

r ^ 1 0 10 ✗ 5 X
lo
2 3 a 8 14 7
6 2
S y
3 8 13
r 5
9 / T
Yy 2
Yz
4 I

7
Chapter 4

Greedy Algo :

An that arrives
algorithm at a solution by '

making of choices , each of which looked

a
sequence
best at the time

Making Change
Problem How can
: a
given amount of
money
be made
w/ the least number of fins with gives demons
Afn put : An of denominations and
integer v
array ,
an

Output the least number

: of coins needed to make

the value V
1. 65 -

1
Coins :{0.05, 0.1 0.25 1,2 }
, ,
0.65 -

0.25
Sum : 1.65
0.4 -

0 .

25
0.15 -

0.10
0.05 -

0.05

knapsack
Fon put objects weight end
n
each object
:
st has a
a valve and a maximum weight W

Output A vector ✗ {✗i. }

.

=
Xz,✗, .
. .
,xn St
-
Of ✗if I -
the total value ✗
, Ytxzvzt txn is
maximised
.
. -

the total
weight
-

✗
txzwzt EW
w
+
, , . . -

Xniwn
Runtime Olnlogln ) ) :
Chapter 4

knapsack lemma
knapsack example
If the objects are
I 1 2 3 4 5
✗g- I selected in
Vi 20 30 66 40 Go decreasing
✗i lo 20 30 40 50
4=1 Viwi , then the greedy
✗2=1
choice finds an optimal
'

2 1.5 2.2 I 1.2

×
; Xg= 0.8
solution
W= 100 Sum __
164

Minimum LMST)
Spanning Tree
Given undirected connected
a
weighted , , Graph G- Cv,r
,

MSTG of
'
an G is sub
a
graph such that :
-

← The vertex set of G' is V

'
-

G is connected
Sum of the
weights of
-

edges in minimum
'
G is

8 7
$ • •

9
4 2
11
y
•
• 14 •
,
10
8
••• -

• •
I z

Lemma : For a Graph G- ( v. E) , split V into A and B .

shortest
edge connected
be A and B
Let { 44 G- E the .

Then there is MST of G

containing {v.v}
an
Chapter 4

Kruskal
5
MST of G 6
Problem: Find an A c E

Approach : Maintain a forest 4

5 3 H
step , add an edge
z
on each p
of min weight that doesn't B D F
create a
cycle z y

Runtime :
Olmlogln)) BGBD.CD.CF.tt#DFEF,.A/B,CE,Ac
1 2 3 4

Prim
start :
-
A-
T
is a set consisting
of
of an
arbitrary vertex
- is an
empty set
edges
Q is all vertices in G- not in A
-

Freq , let :

of any edge and

miniweightlv
vertex in A
) : min
weight v a

closest 14 : vertex u in A- s.twtlv.ir)= minweightlv)

Iteration :

Select the vertice ✗ st min minimum

-

weight 1×1 is
-
Add ✗ to A and remove it from Q
-
add
{× ,
closest 1×1 } to T
-
update closest and minmeight for all vertices
not in A. connected to ✗

Runtime : Olmloglnl )
Chapter 4

Prim -

example
6 5 b
A B C D
1
>
1 23 5 5 7
2 46

F- 2
43
5
F C- H
,

:
Chapter 5 General structure
B. Structure the optimal
solution
2. Write a recursion for
the optimal solution
3. Solve the recurrence

Matrix Multiplication
Problem : Find the optimal order to multiply a set
of matrices
together matrices Al Az Az
F- input :
A set of n
, ,
An where
, , . . .

A, is Po XP,
,
Az is P, Pz ,
✗
,
Antis
. .
Pn i ✗ Pn
. .
-
-

Output : the number of steps needed to multiply all

matrices optimally

,j)={iI¥jk
0 if i=j
mli
mliikltmlktl 's ) -1 Pi
, Pkpj ) lficjsn ,

Runtime : 01ns)
-

Longest common
subsequence (Lcs)
Problem : Find the LCS between two sets

Input : 2
sequences
✗ and Y
between ✗ and Y
Output : the LCS

Runtime : 01m .nl

{[a×(
i = 0 or j = 0

Cli , j ) = 1 + Cli -

1, j -

1) i >o ,j >o and ✗

i=Yj
efi,j -11 di I ,j )) i > -0
,j > 0
-

and ✗ i =/
,
Yj
Chapter 5

Floyd -
War shall

Problem : Find the shortest path between all pairs of

vertices in a
weighted directed,
graph C-

ftp.nput : Graph G- IV. E) ,

G- is directed and
edges are weighted
For Asian ,: dist ( ii. 01--0

For A fish If Jen

, , izj :

,j,o)={
tliii ) if
distli edge fisj ) exists
✗
else
For Asian , 1£ js.in Is Ken :
,

distci,j,k)=min{
distli.j.k-H.distli.kik-iltdistlkis.tl-11)
Runtime: 01h31
Chapter 6

A decision problem is a
problem for which

the answer is Yes or no

class p Class NP ( non-deterministic)

NP if is in
A problem A :

set of all problems For given input Ii

-

that can be solved the answer to ACI) is

yes
then 1- a proof es .t
in time
:

polynomial -
C is polynomial to the
length of I
-
In Polynomial time ,
Problems in P we
verify that
can

C is a correct proof
graph bipartite
?
-
Is a

Is connected ? that ACII Yes

graph
-

-
-
a

1-1am Cycle-
Problem
Input Undirected graph
:
G- =
IV. E)

Question : Does G- contain a hamiltonian cycle?

• • Unknown
if in P
•
•

•
•
Chapter 6

Traveling Salesman Problem LTSP) Problem

Input :
A. complete directed graph
-

G- LYE)
,
where
each
edge is
weighed
An
- K
integer
Question : Does G contain a hamiltonian cycle
of total at most
weight K?

Unknown if in p

Subset Sum - Problem

Input :

A of
5={5,756/23,2902}
set
integers g
-

f- 2963
-
an
integer K
Yes ! { 5,5612902}
'
S
Question : Is
-

there a subset g)
of S Unknown if in P
s.tt#s.x--t ?
Chapter 6
Definition of P
A
The
language L is in P if 7 an
algo
and a
constant cat sit for any input ✗

-
✗ EL Alex)= Yes
-
The runtime of Atx ) is olnc) where n

is the
length of ✗

Definition of Np ( Non -
deterministic Polynomial)
The language L is in NP if F- an
algor
and a Constante 21 s -
t for any input ✗

✗ EL ⇐ 1- a certificate yet
-
Iyl =
011×14
-
Vlx, y ) =
Yes
- VCX /
y)
runtime is polynomial
of
to the
length ✗

Theorem Language
NP The
language
of

P c- NP a Problem
decision
p is the set of all
inputs for which
the answer is

yes or no
Chapter 6
Reductions Famous 3 Properties
'
L and L
For 2 languages
J
.

1. f maps inputs for Ltoli

L is polynomial-time
we
say
reducible to L
'
if -7 function
2. It inputs ✗ for L
a '

✗ c- L⇐> flxl
C- L

f which satisfies the

3. inputs ✗ for L ,
famous three properties '
flx) is 0cm for
Notation : L
•

CZI
Lsp

CLIQUE Problem
K=4
Input :
• • •

graph
•
,

- An undirected C-
An integer K •
••
-

Question : Does G contain •

a clique of size K

INDEP -
SET
k =3
Problem
In plot : ⑨ •

\
- An undirected graph C-
- An integer K
⑨
Question Does : C- contain
an independent set of size
K
•
⑨
INDEP -
SET
Ep CLIQUE
Chapter 6 Theorem
If LEPL '
and L' EP
then LEP

VERTEX COVER - Problem

k=4
Input :
undirected ⑨
graph G
•
An
•
-

-
An int k •

Question :
Does ' G contain ⑧ ⑨ ⑨
a vertex cover of size K
-

CLIQUE ftp.VERTEX -

COVER

3. SAT
µ%npUt ?
10=1> ✗ vyv> 2) nlxv > yvz )A(xvyvz't
y
"

-
A boolean formula of the
Y ✗

form 10=9^9 ^ . . -
Mcm
'

Clause
↳ ci-ljvli.ve , literal
7✗ 72 TY -2

Question : Is at satisfiable? ×

- Does there exist a truth

value for the literals Y z

s.to/--l3SATfp.INDEP- SET
Chapter 6
NP-Hard NP-Complete
A L is NP-complete
A
language
L is NP-Hard
if
language
:
if Y L' ENP, L'
Ep L -
LENP
t L'c- NP L' fpl
Not covered
-

Theorem Million $
Then
Assume L is NP-complete Question
LEP < P=NP

Theorem

}
L NP-complete
is
'
L NP-complete
'
is
Lfp L
L' GNP

'
show L is NP-complete
1. Show L' ENP
2. Look for "
similar
"
problem L that
is NP Complete
-

'
3. Show Lfpl
Chapter 6
CIRCUIT -
SAT
Input :

Directed acyclic graph AND/OR Gates have

- -
-

whose vertices in
degree
are -

2
boolean
gates NOT Gates
-

have
Known input gates
-

in
have in
degree 0 and
-

degree I
are TRUE or FALSE
-
There is one output
-
Unknown input gates gate
is 0
,
whose out
degree
have
and
in
are
degree
labelled ?
0

Question : Is possible to truth value

it
assign
-

to each Inputs .t is TRUE

unknown the output

complete
'

CIRCUIT SAT is NP -
-

CIRCUIT
-
SAT fp SAT 3
Chapter 7
Theorem
In the decision tree moder
,

of numbers
sorting array n
an

takes rent
ogcnll time

Theorem
In the decision tree model ,

in sorted
finding the position of × a

array A -11 .n] takes D. llogcnl) time

. .

Theorem
Icn) time to find the
max
It takes
element in an A- [ 1 n]
array
. - •

Theorem
It takes h1n1 steps to compute
the median of an unsorted array AII . . ]
n