Greedy Strategy And Dynamic

Programming

1
 A short list of categories so far covered

 Simple recursive algorithms

 Divide and conquer algorithms

 Greedy algorithms

 Dynamic Programming

2
2
 Contents
 Knapsack problem

 Huffman code generation algorithm

 Job Sequencing with Deadlines

3
3
 Optimization problems
 An optimization problem is one in which you want
to find, not just a solution, but the best solution
 A “greedy algorithm” sometimes works well for
optimization problems
 A greedy algorithm works in phases. At each
phase:
 You take the best you can get right now, without regard
for future consequences
 You hope that by choosing a local optimum at each
step, you will end up at a global optimum

4
4
 Introduction : Greedy Method
 A greedy algorithm for an optimization problem
always makes the choice that looks best at the
moment and adds it to the current subsolution.

 Final output is an optimal solution.

 Greedy algorithms don’t always yield optimal

solutions but, when they do, they’re usually the
simplest and most efficient algorithms available.

5
 Introduction : Greedy Method
 It is used to solve problems that have ‘n’ inputs and
require us to obtain a subset that satisfies some
constraints.
 Any subset that satisfied the constraints is called as a
feasible solution.
 We need to find the optimum feasible solution i.e. the
feasible solution that optimizes the given objective
functions.
 It works in stages. At each stage, At each stage a decision
is made regarding weather or particular input is in the
optimum solution.
6
6
 Introduction : Greedy Method
 The greedy method suggests that one can divide the algorithm
that works in stages. Considering one input at a time.
 At each stage a decision is made regarding weather or
particular input is in the optimum solution.
 For this purpose all the inputs must be arranged in a
particular order by using a selection process.
 If the inclusion of next input in to the partially constructed
solution will result in an infeasible solution then the input will
not be considered and will not be added to partially
constructed set otherwise it is added.
 CHANGE-MAKING PROBLEM (coin changing)

7
7
 Coin Changing
An optimal solution to the coin changing problem is the minimum number of
coins whose total value equals a specified amount. For example what is the
minimum number of coins (current U.S. mint) needed to total 83 cents.
Objective function: Minimize number of coins returned.
Greedy solution: Always return the largest coin you can
A Greedy Algorithm for
Coin Changing
83 - 25 = 58
58 - 25 = 33 1. Set remval=initial_value
33 - 25 = 8 2. Choose largest coin that
8 - 5 = 3 is less than remval.
3 - 1 = 2
3. Add coin to set of coins
2 - 1 = 1 and set
remval:=revamal-coin_value
1 - 1 = 0
4. repeat Steps 2 and 3
8
until remval = 0; 8
 Greedy Method
 Feasible Solution:-
Any subset of the solutions that satisfies the constraints of the problem
is known as a feasible solution.
 Optimal Solution:-

The feasible solution that maximizes or minimizes the given objective

function is an optimal solution.
Example : MST

9
 Control Abstraction : Greedy Method
Algorithm Greedy (a,n)
//a[1:n] contains the ‘n’ inputs
{
Solution: = 0; //initialize the solution
for i: = 1 to n do
{
x: = select (a);
if Feasible (solution, x) then
solution: = Union (solution, x);
}
return solution;
}

 Select is a function which is used to select an input from the set.

 Feasible is a function which verifies the constraints and determines
weather the resultant solution is feasible or not.
 Union is a function which is used to add elements to the partially
constructed set. 10
 Knapsack Problem
 Given :

A knapsack with capacity ‘m’

A set of ‘n’ objects with each item i having

pi - a positive benefit

wi - a positive weight
 Goal: Choose items with maximum total benefit but with weight at most m.

 If we are allowed to take fractional amounts, then this is the fractional knapsack problem.

Let xi denote the amount we take of item i , 0≤ xi ≤ 1
 Objective: maximize ∑ pi xi
1≤i≤n

 Constraint: ∑ wi xi ≤ m
1≤i≤n

and 0 ≤ xi ≤ 1 , 1 ≤ i ≤ n

11
 Knapsack Problem : Example
Ex: - Consider following instance of Knapsack Problem with 3 objects whose
profits and weights are defined as
(P1, P2, P3) = ( 25, 24, 15 ) (W1, W2, W3) = ( 18, 15, 10 )
n=3 Knapsack Capacity m=20
Determine the optimum strategy for placing the objects in to the knapsack.

The four feasible solutions are :

(1) Greedy about profit: -
Profit Weight
(x1, x2, x3) Order Order ∑ xiwi <=20 ∑ xipi
(p1,p2,p3) (w1,w2,w3)
(1, 2/15, 0) (25,24,15) (18, 15, 10 ) 18 x 1+(2/15) x 15 = 20 25 x 1+ (2/15) x 24 = 28.2

(2) Greedy about weight: -

Profit Weight
(x1, x2, x3) Order Order ∑ xiwi <=20 ∑ xipi
(p3,p2,p1) (w3,w2,w1) 18 x 0+(2/3) x 15+10 =
(0, 2/3, 1) (15, 24,25) (10,15,18 ) 20 25 x0+ (2/3) x 24 +25= 31
12
12
(3) Greedy about profit / unit weight: -

(x1, x2, x3) ∑ xiwi <=20 ∑ xipi

18 x 0+1 x 15+(1/2)x10 = 25 x 0+ 1 x 24+(1/2)x15 =
(0, 1, ½ ) 20 31.5

(4) If an additional constraint of including each and every object is placed then the greedy strategy
could be

(½, ⅓, ¼ ) ∑ xiwi = ½ x 18+⅓ x15+ ¼ x10 = 16. 5

∑ xipi =½ x 25+⅓ x24+ ¼ x15 = 12.5+8+3.75 = 24.25

(x1, x2, x3) ∑ xiwi <=20 ∑ xipi

(½, ⅓, ¼ ) 16.5 24.25

13
 Algorithm : Greedy Knapsack
Algorithm Greedy knapsack (m, n)
//p (1:n) and w(1:n) contain the profits and weights resp. of the n objects
//ordered such that p(i)/W(i)  p(i+1)/w (i+1). M is the knapsack size and x (1:n)
// is the solution vector
{
for i: = 1 to n do x(i) = 0.0i// initialize x O(n)
u : = m;
for i: = 1 to n do O(n)
{
if (w(i) > u) then break;
x(i): = 1.0;
u: = u-w(i);
}
if (i≤n) then x(i): = u/w(i);
}
 Analysis: - If. we do not consider the time considered for sorting the inputs then the
complexity will be O(n).

14
 Knapsack : More examples

1) Greedy approach- consider the following instances of knapsack problem

n=5, w=100, W (10, 20, 30, 40, 50) , V(20, 30, 66, 40, 60 ), find the
optimal solution.
2)

3) Let us consider that the capacity of the knapsack W =60 and the list of provided
items are shown in the following table −
Item A B C D
Profit 280 100 120 120
Weight 40 10 20 24
15
 Huffman Coding

 Huffman codes can be used to compress

information
 JPEGs do use Huffman as part of their compression
process

 The basic idea is to store the more frequently

occurring characters using fewer bits and less
frequently occurring characters using more bits
 On average this should decrease the filesize (usually ½)

16
 Huffman Coding

 As an example, lets take the string:

“duke blue devils”
 We first to a frequency count of the characters:

e:3, d:2, u:2, l:2, space:2, k:1, b:1, v:1, i:1, s:1
 Next we use a Greedy algorithm to build up a Huffman
Tree
 We start with nodes for each character

e,3 d,2 u,2 l,2 sp,2 k,1 b,1 v,1 i,1 s,1
17
 Huffman Coding

 We then pick the nodes with the smallest frequency and

combine them together to form a new node
 The selection of these nodes is the Greedy part

 The two selected nodes are removed from the set, but
replace by the combined node

 This continues until we have only 1 node left in the set

18
 Huffman Coding

e,3 d,2 u,2 l,2 sp,2 k,1 b,1 v,1 i,1 s,1

19
 Huffman Coding

e,3 d,2 u,2 l,2 sp,2 k,1 b,1 v,1 2

i,1 s,1

20
 Huffman Coding

e,3 d,2 u,2 l,2 sp,2 k,1 2 2

b,1 v,1 i,1 s,1

21
 Huffman Coding

e,3 d,2 u,2 l,2 sp,2 3 2

k,1 2 i,1 s,1

b,1 v,1

22
 Huffman Coding

e,3 d,2 u,2 4 3 2

l,2 sp,2 k,1 2 i,1 s,1

b,1 v,1

23
 Huffman Coding

e,3 4 4 3 2

d,2 u,2 l,2 sp,2 k,1 2 i,1 s,1

b,1 v,1

24
 Huffman Coding

e,3 4 4 5

d,2 u,2 l,2 sp,2 2 3

i,1 s,1 k,1 2

b,1 v,1

25
 Huffman Coding

7 4 5

e,3 4 l,2 sp,2 2 3

d,2 u,2 i,1 s,1 k,1 2

b,1 v,1

26
 Huffman Coding

7 9

e,3 4 4 5

d,2 u,2 l,2 sp,2 2 3

i,1 s,1 k,1 2

b,1 v,1

27
 Huffman Coding

16

7 9

e,3 4 4 5

d,2 u,2 l,2 sp,2 2 3

i,1 s,1 k,1 2

b,1 v,1
28
 Huffman Coding
 Now we assign codes to the tree by placing a 0 on every
left branch and a 1 on every right branch

 A traversal of the tree from root to leaf give the

Huffman code for that particular leaf character

 Note that no code is the prefix of another code

29
 Huffman Coding
e 00
16 d 010
u 011
7 9 l 100
sp 101
e,3 4 4 5 i 1100
s 1101
d,2 u,2 l,2 sp,2 2 3 k 1110
b 11110
i,1 s,1 k,1 2 v 11111

b,1 v,1
* Using variable length encoding

30
 Huffman Encoding cormen
HUFFMAN (C )
1. n= |C|
//initializes the min-priority queue Q with the characters in C
2. Q = C O(n)
//loop extracts the two nodes x and y of lowest frequency from the queue

3. for i = 1 to n – 1 O(n log n)

//replacing them in the queue with a new node z representing their merger
4. allocate a new node z
5. z.left = x = EXTRACT-MIN(Q) O(log n)
6. z.right = y = EXTRACT-MIN(Q)
7. z.freq = x:freq + y:freq
8. INSERT (Q,z)
//After n-1 mergers, line 9 returns the one node left in the queue,which is the root of the code tree
9. return EXTRACT-MIN(Q) // return the root of the tree
Analysis : Must sort n values before making n choices. Therefore, Huffman is O(n log n) + O(n)
= O(n log n)

31
31
 Huffman Coding
 These codes are then used to encode the string

 Thus, “duke blue devils” turns into:

010 011 1110 00 101 11110 100 011 00 101 010 00 11111 1100 100 1101

 When grouped into 8-bit bytes:

01001111 10001011 11101000 11001010 10001111 11100100 1101xxxx

 Thus it takes 7 bytes of space compared to 16 characters * 1

byte/char = 16 bytes uncompressed

33
 Examples : Huffman Coding
 Solve the following using Huffman's code generation
algorithm

Huffman Coding Algorithm Example

Construct a Huffman tree by using these nodes
Value A B C D E F
Freq 5 25 7 15 4 12
uenc
y
34
 Other greedy algorithms
 Dijkstra’s algorithm for finding the shortest path in a
graph
 Always takes the shortest edge connecting a known node to
an unknown node
 Kruskal’s algorithm for finding a minimum-cost
spanning tree
 Always tries the lowest-cost remaining edge
 Prim’s algorithm for finding a minimum-cost spanning
tree
 Always takes the lowest-cost edge between nodes in the
spanning tree and nodes not yet in the spanning tree

35
35
 JOB SEQUENCING WITH
DEADLINES
The problem is stated as below.
• There are n jobs to be processed on a machine.
• Each job i has a deadline di≥ 0 and profit pi≥0 .
• Pi is earned iff the job is completed by its deadline.
• The job is completed if it is processed on a machine
for unit time.
• Only one machine is available for processing jobs.
• Only one job is processed at a time on the machine.
36
36
 JOB SEQUENCING WITH
DEADLINES (contd..)
• A feasible solution is a subset of jobs J such that
each job is completed by its deadline.
• An optimal solution is a feasible solution with
maximum profit value.
• NO LATER THAN THEIR RESPECTIVE
DEADLINES.
Example : Let n = 4, (p1,p2,p3,p4) = (100,10,15,27),
(d1,d2,d3,d4) = (2,1,2,1)

37
37
 JOB SEQUENCING WITH
DEADLINES (contd..)
Example : Let n = 4, (p1,p2,p3,p4) = (100,10,15,27),
(d1,d2,d3,d4) = (2,1,2,1)
Sr.No. Feasible Processing Profit value
Solution Sequence
(i) (1,2) (2,1) 110
(ii) (1,3) (1,3) or (3,1) 115
(iii) (1,4) (4,1) 127 is the optimal one
(iv) (2,3) (2,3) 25
(v) (3,4) (4,3) 42
(vi) (1) (1) 100
(vii) (2) (2) 10
(viii) (3) (3) 15
38
(ix) (4) (4) 27 38
 Example Explanation
• Consider the jobs in the non increasing order of
profits subject to the constraint that the resulting
job sequence J is a feasible solution.
• In the example considered before, the non-
increasing profit vector is
(100 27 15 10) (2 1 2 1) p 1 p4
p3 p2 d1 d4 d3 d2

39
39
GREEDY ALGORITHM TO OBTAIN
AN OPTIMAL SOLUTION (Contd..)

(100 27 15 10) (2 1 2 1)
p1 p 4 p3 p2 d1 d4 d3 d2
J = { 1} is a feasible one
J = { 1, 4} is a feasible one with processing
sequence ( 4,1)
J = { 1, 3, 4} is not feasible
J = { 1, 2, 4} is not feasible
J = { 1, 4} is optimal
40
40
 GREEDY ALGORITHM TO OBTAIN
AN OPTIMAL SOLUTION (Contd..)

J(r+1)i ; k k+1
// i is inserted at position r+1 //
// and total jobs in J are increased by one //
repeat
end JS

41
41
Algorithm for Job sequencing with
Deadlines

42
 COMPLEXITY ANALYSIS OF JS
ALGORITHM
• Let n be the number of jobs and s be the number of
jobs included in the solution.
• The loop between lines 4-15 (the for-loop) is
iterated (n-1)times.
• Each iteration takes O(k) where k is the number of
existing jobs.
 The time needed by the algorithm is 0(sn) s  n so
the worst case time is 0(n2).
If di = n - i+1 1  i  n, JS takes θ(n2) time
D and J need θ(s) amount of space.
43
43
Examples :Job Sequencing Problems:
1) Write a greedy algorithm for sequencing unit time jobs with deadlines
and profits. Using this algorithm, find the optimal solutions when
n=5,(p1,p2,p3,p4,p5)=(20,15,10,5,1) and (d1,d2,d3,d4,d5)=(2,2,1,3,3).

2) Find the correct sequence for jobs using following instances,

JobID Deadline Profit
1 4 20
2 1 10
3 1 40
4 1 30
3)Find the correct sequence for jobs using following instances,
JobID Deadline Profit
1 2 100
2 1 19
3 2 27
4 1 25
5 3 15
4)Find the optimum job sequence for the following problem.
n=7 P7 J1 j2 j3 j4 j5 j6 j7
(P1, P2, P3, P4, P5, P6) = (3, 5, 10, 18, 1, 6, 30)
, d4, d5, d6) = (1,3, 4, 3, 2, 1, 2)
Feasible soln. Processing Sequence Value
(1, 2, 3, 4) (1,2,4,3) or (1,4,2,3) 3+5+10+18= 36

44
The End

45
45