Scribd
Upload
15 views

Dynamic Programming Lecture 2

The document discusses dynamic programming and how it can be used to efficiently solve problems with optimal substructure by storing results of subproblems to avoid recomputing them. It provides examples of using dynamic programming to calculate Fibonacci numbers and binomial coefficients in linear time by storing intermediate results rather than recursively calculating them.

Uploaded by

debajyotichakraborty92
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
15 views

Dynamic Programming Lecture 2

The document discusses dynamic programming and how it can be used to efficiently solve problems with optimal substructure by storing results of subproblems to avoid recomputing them. It provides examples of using dynamic programming to calculate Fibonacci numbers and binomial coefficients in linear time by storing intermediate results rather than recursively calculating them.

Uploaded by

debajyotichakraborty92
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 27

Lecture 16:

Introduction to Dynamic Programming


Steven Skiena

Department of Computer Science


State University of New York
Stony Brook, NY 11794–4400

http://www.cs.stonybrook.edu/˜skiena
Topic: Problem of the Day
Problem of the Day
Multisets are allowed to have repeated elements. A multiset
of n items may thus have fewer than n! distinct permutations.
For example, {1, 1, 2, 2} has only six different permutations:
{1, 1, 2, 2}, {1, 2, 1, 2}, {1, 2, 2, 1}, {2, 1, 1, 2}, {2, 1, 2, 1},
and {2, 2, 1, 1}. Design and implement an efficient algorithm
for constructing all permutations of a multiset.
Questions?
Topic: Introduction to Dynamic Programming
Dynamic Programming
Dynamic programming is a very powerful, general tool for
solving optimization problems on left-right-ordered items
such as character strings.
Once understood it is relatively easy to apply, it looks like
magic until you have seen enough examples.
Floyd’s all-pairs shortest-path algorithm was an example of
dynamic programming.
Greedy vs. Exhaustive Search
Greedy algorithms focus on making the best local choice at
each decision point. In the absence of a correctness proof
such greedy algorithms are very likely to fail.
Dynamic programming gives us a way to design custom
algorithms which systematically search all possibilities (thus
guaranteeing correctness) while storing results to avoid
recomputing (thus providing efficiency).
Recurrence Relations
A recurrence relation is an equation which is defined in terms
of itself. They are useful because many natural functions are
easily expressed as recurrences:
Polynomials: an = an−1 + 1, a1 = 1 −→ an = n
Exponentials: an = 2an−1, a1 = 2 −→ an = 2n
Weird: an = nan−1, a1 = 1 −→ an = n!
Computer programs can easily evaluate the value of a given
recurrence even without the existence of a nice closed form.
Questions?
Topic: Fibonacci Numbers
Computing Fibonacci Numbers

Fn = Fn−1 + Fn−2, F0 = 0, F1 = 1
Implementing this as a recursive procedure is easy, but slow
because we keep calculating the same value over and over.
long fib_r(int n) {
if (n == 0) {
return(0);
}

if (n == 1) {
return(1);
}

return(fib_r(n-1) + fib_r(n-2));
}
How Slow?
√
Fn+1/Fn ≈ φ = (1 + 5)/2 ≈ 1.61803
F(6)=8

F(5) F(4)

F(4) F(3) F(3) F(2)

F(2) F(1) F(2) F(1) F(1) F(0)


F(3) F(2)

F(2) F(1) F(1) F(0) F(1) F(0) F(1) F(0)

F(1) F(0)

Thus Fn ≈ 1.6n. Since our recursion tree has 0 and 1 as


leaves, computing Fn requires ≈ 1.6n calls!
What about Memoization?
We can explicitly cache calls after computing results to avoid
recomputation:
#define MAXN 92 /* largest n for which F(n) fits in a long */
#define UNKNOWN -1 /* contents denote an empty cell */
long f[MAXN+1]; /* array for caching fib values */

long fib_c(int n) {
if (f[n] == UNKNOWN) {
f[n] = fib_c(n-1) + fib_c(n-2);
}

return(f[n]);
}
long fib_c_driver(int n) {
int i; /* counter */

f[0] = 0;
f[1] = 1;

for (i = 2; i <= n; i++) {


f[i] = UNKNOWN;
}

return(fib_c(n));
}
What about Dynamic Programming?
We can calculate Fn in linear time by storing small values:
F0 = 0
F1 = 1
For i = 1 to n
Fi = Fi−1 + Fi−2
Moral: we traded space for time.
Fibonacci by Dynamic Programming

long fib_dp(int n) {
int i; /* counter */
long f[MAXN+1]; /* array for caching values */

f[0] = 0;
f[1] = 1;

for (i = 2; i <= n; i++) {


f[i] = f[i-1] + f[i-2];
}

return(f[n]);
}
Why I Love Dynamic Programming
Dynamic programming is a technique for efficiently comput-
ing recurrences by storing partial results.
Once you understand dynamic programming, it is usually
easier to reinvent certain algorithms than try to look them up!
I have found dynamic programming to be one of the most
useful algorithmic techniques in practice:
• Morphing in computer graphics.
• Data compression for high density bar codes.
• Designing genes to avoid or contain specified patterns.
Questions?
Topic: Binomial Coefficients
Avoiding Recomputation by Storing Results
The trick to dynamic programmming is to see that the
naive recursive algorithm repeatedly computes the same
subproblems over again, so storing the answers in a table
instead of recomputing leads to an efficient algorithm.
We first hunt for a correct recursive algorithm, then we try to
speed it up by using a results matrix.
Binomial Coefficients
The most important class of counting numbers are the
binomial coefficients, where ( ) counts the number of ways
n
k

to choose k things out of n possibilities.


• Committees – How many ways are there to form a k-
member committee from n people? By definition, ( ).
n
k

• Paths Across a Grid – How many ways are there to travel


from the upper-left corner of an n × m grid to the lower-
right corner by walking only down and to the right? Every
path must consist of n + m steps, n downward and m to
the right, so there are ( ) such sets/paths.
n+m
n
Computing Binomial Coefficients
Since ( ) = n!/((n − k)!k!), in principle you can compute
n
k

them straight from factorials.


However, intermediate calculations can easily cause arith-
metic overflow even when the final coefficient fits comfort-
ably within an integer.
Pascal’s Triangle
No doubt you played with this arrangement of numbers in
high school. Each number is the sum of the two numbers
directly above it:
Pascal’s Recurrence
A more stable way to compute binomial coefficients is using
the recurrence relation implicit in the construction of Pascal’s
triangle, namely, that

()=( )+( )
n
k
n−1
k−1
n−1
k

It works because the nth element either appears or does not


appear in one of the ( ) subsets of k elements.
n
k
Basis Case
No recurrence is complete without basis cases.
How many ways are there to choose 0 things from a set?
Exactly one, the empty set.
The right term of the sum drives us up to ( ). How many ways
k
k

are there to choose k things from a k-element set? Exactly


one, the complete set.
Binomial Coefficients Implementation

long binomial_coefficient(int n, int k) {


int i, j; /* counters */
long bc[MAXN+1][MAXN+1]; /* binomial coefficient table */

for (i = 0; i <= n; i++) {


bc[i][0] = 1;
}

for (j = 0; j <= n; j++) {


bc[j][j] = 1;
}

for (i = 2; i <= n; i++) {


for (j = 1; j < i; j++) {
bc[i][j] = bc[i-1][j-1] + bc[i-1][j];
}
}

return(bc[n][k]);
}
Questions?

You might also like