Thanks to visit codestin.com
Credit goes to www.scribd.com

Scribd
Upload
20 views4 pages

33 Class

The document discusses methods for finding the convex hull of a set of points, including Graham's scan and Jarvis's March. It outlines the time complexities of these algorithms, with Graham's scan being O(n log n) and Jarvis's March being O(nh), where h is the number of convex hull vertices. The document also includes details on the steps involved in each algorithm and their respective correctness.

Uploaded by

Kevin
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
20 views4 pages

33 Class

The document discusses methods for finding the convex hull of a set of points, including Graham's scan and Jarvis's March. It outlines the time complexities of these algorithms, with Graham's scan being O(n log n) and Jarvis's March being O(nh), where h is the number of convex hull vertices. The document also includes details on the steps involved in each algorithm and their respective correctness.

Uploaded by

Kevin
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 4

for i = 3 to n do

------------------------------- 33 ------------------------------ Hull(p0, p1, ..., pi) 33-2


Computational Geometry

P2
33.3 Finding the convex hull P1
P0
The convex hull of a set Q of points is the smallest
convex polygon P for which each point in Q is Stack
either on the boundary of P or in its interior. Hull(p0, p1, p2)

P3
P1
P0

Hull(p0, p1, p2, p3)

* A simple O(n3) solution: if ab is a convex edges,


all other points are at the same side. P4
Graham’s scan (Output the vertices in
counterclockwise order.) P3
always make P1
counter-
clockwise a left-turn P0
33-1a
Hull(p0, p1, ..., p4)
33-3 33-4
P8
P7
P6
P5 P5
P3 P3
P1 P1
P0 P0
Hull(p0, p1, ..., p5)

P6 P9
P5 P6
P3 P5
P1 P3
P0 P1
Hull(p0, p1, ..., p6) P0

P7
P6
P5 P10
P3 P3
P1 P1
P0 P0
Hull(p0, p1, ..., p7)
33-5 33-6

O(n)

O(nlg n)

(How? See 33-1b)


function compare(p0, a, b)
0: a  b
1: a > b
O(n)
multi-pop + push (amortized)

Top Pi

Next-to-top
(How? See 33-1a)
Time complexity:

Line 1: O(n) find P0


Line 2: O(nlg n) P1, P2, P3, ..., Pm (m  n-1)
Line 3~6: O(1)
Line 8~9: O(n) Amortized
(at most n PUSH  at most n POP)
Line 10: O(n) * if the sorted order * take any as p0:
O(nlg n) is given : O(n) 33-6x

* Correctness: See textbook.


The end!
33-7 33-8
Jarvis’s March (using a technique known as 33-7x

package wrapping or gift wrapping) left chain right chain


(Find a correct edge each time !)
Step 1: Find the lowest point x and the highest
point y. Smallest 
Step 2: Compute the right chain negative x-axis
(p0=x, p1, …, pk=y).
Step 3: Compute the left chain
(pk=y, pk+1, …, ph=x).

33-8x
* In O(1) time, we can compare the polar angles of
two points. (How ??? See Section 33.1) 0: a  b
function compare(p0, a, b): 1:  > 
a b

* In O(n) time, we can determine the point with Smallest 


smallest (or largest) polar angle with respect to
a given point. (How? See 33-1b)

* Since each computation of pi take O(n) time,


T(n)=O(nh).
# of convex hull vertices
Homework: None. conclude

* In the worst case, h=n and thus T(n)=O(n2).

* Jarvis’s march is better than Graham’s scan if


h=o(lg n).

You might also like