Probabilistic algorithms for extreme point identification

Shafiu Jibrin, Arnon Boneh, Richard J. Caron

Research output: Contribution to journalArticle

4 Citations (Scopus)

Abstract

For a finite set of distinct points S = {pi ε l} in ℝd there exists Î ⊆ I such that all points in Ŝ = {pi i ε Î} are extreme points and conv(Ŝ) = conv(S). Since a point pk ε S is extreme if and only if the inequality pTk y ≤ 1 is necessary with respect to the representation {y ε ℝd{pipe} pTi y≤ 1, i ε I} of the polar dual SΔ of S, Ŝ can be found by classifying the inequalities in the representation as necessary or redundant. Thus, the problem of finding Ŝ is polynomial. This paper shows the advantage of using probabilistic hit-and-run algorithms applied to the polar dual for the quick identification of points in Ŝ and shows how, in an application to a certain cluster analysis problem, it can be used to also identify points in S/Ŝ. Further, it shows that the hit-and-run variant known as Stand-and-Hit provides approximations for the exterior solid angles at the extreme points of S.

Original languageEnglish (US)
Pages (from-to)131-142
Number of pages12
JournalJournal of Interdisciplinary Mathematics
Volume10
Issue number1
DOIs
StatePublished - 2007

Fingerprint

Probabilistic Algorithms
Cluster analysis
Extreme Points
Pipe
Polynomials
Hits
Solid angle
Necessary
Cluster Analysis
Pi
Finite Set
Extremes
If and only if
Distinct
Polynomial
Approximation

Keywords

  • Convex hull
  • Extreme points
  • Linear inequalities
  • Probabilistic method
  • Redundancy
  • Solid angles

ASJC Scopus subject areas

  • Analysis
  • Applied Mathematics

Cite this

Probabilistic algorithms for extreme point identification. / Jibrin, Shafiu; Boneh, Arnon; Caron, Richard J.

In: Journal of Interdisciplinary Mathematics, Vol. 10, No. 1, 2007, p. 131-142.

Research output: Contribution to journalArticle

Jibrin, Shafiu ; Boneh, Arnon ; Caron, Richard J. / Probabilistic algorithms for extreme point identification. In: Journal of Interdisciplinary Mathematics. 2007 ; Vol. 10, No. 1. pp. 131-142.
@article{c1c7fdceb5094130ac58cfedca60d837,
title = "Probabilistic algorithms for extreme point identification",
abstract = "For a finite set of distinct points S = {pi ε l} in ℝd there exists {\^I} ⊆ I such that all points in Ŝ = {pi i ε {\^I}} are extreme points and conv(Ŝ) = conv(S). Since a point pk ε S is extreme if and only if the inequality pTk y ≤ 1 is necessary with respect to the representation {y ε ℝd{pipe} pTi y≤ 1, i ε I} of the polar dual SΔ of S, Ŝ can be found by classifying the inequalities in the representation as necessary or redundant. Thus, the problem of finding Ŝ is polynomial. This paper shows the advantage of using probabilistic hit-and-run algorithms applied to the polar dual for the quick identification of points in Ŝ and shows how, in an application to a certain cluster analysis problem, it can be used to also identify points in S/Ŝ. Further, it shows that the hit-and-run variant known as Stand-and-Hit provides approximations for the exterior solid angles at the extreme points of S.",
keywords = "Convex hull, Extreme points, Linear inequalities, Probabilistic method, Redundancy, Solid angles",
author = "Shafiu Jibrin and Arnon Boneh and Caron, {Richard J.}",
year = "2007",
doi = "10.1080/09720502.2007.10700478",
language = "English (US)",
volume = "10",
pages = "131--142",
journal = "Journal of Interdisciplinary Mathematics",
issn = "0972-0502",
publisher = "TARU Publications",
number = "1",

}

TY - JOUR

T1 - Probabilistic algorithms for extreme point identification

AU - Jibrin, Shafiu

AU - Boneh, Arnon

AU - Caron, Richard J.

PY - 2007

Y1 - 2007

N2 - For a finite set of distinct points S = {pi ε l} in ℝd there exists Î ⊆ I such that all points in Ŝ = {pi i ε Î} are extreme points and conv(Ŝ) = conv(S). Since a point pk ε S is extreme if and only if the inequality pTk y ≤ 1 is necessary with respect to the representation {y ε ℝd{pipe} pTi y≤ 1, i ε I} of the polar dual SΔ of S, Ŝ can be found by classifying the inequalities in the representation as necessary or redundant. Thus, the problem of finding Ŝ is polynomial. This paper shows the advantage of using probabilistic hit-and-run algorithms applied to the polar dual for the quick identification of points in Ŝ and shows how, in an application to a certain cluster analysis problem, it can be used to also identify points in S/Ŝ. Further, it shows that the hit-and-run variant known as Stand-and-Hit provides approximations for the exterior solid angles at the extreme points of S.

AB - For a finite set of distinct points S = {pi ε l} in ℝd there exists Î ⊆ I such that all points in Ŝ = {pi i ε Î} are extreme points and conv(Ŝ) = conv(S). Since a point pk ε S is extreme if and only if the inequality pTk y ≤ 1 is necessary with respect to the representation {y ε ℝd{pipe} pTi y≤ 1, i ε I} of the polar dual SΔ of S, Ŝ can be found by classifying the inequalities in the representation as necessary or redundant. Thus, the problem of finding Ŝ is polynomial. This paper shows the advantage of using probabilistic hit-and-run algorithms applied to the polar dual for the quick identification of points in Ŝ and shows how, in an application to a certain cluster analysis problem, it can be used to also identify points in S/Ŝ. Further, it shows that the hit-and-run variant known as Stand-and-Hit provides approximations for the exterior solid angles at the extreme points of S.

KW - Convex hull

KW - Extreme points

KW - Linear inequalities

KW - Probabilistic method

KW - Redundancy

KW - Solid angles

UR - http://www.scopus.com/inward/record.url?scp=56049096100&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=56049096100&partnerID=8YFLogxK

U2 - 10.1080/09720502.2007.10700478

DO - 10.1080/09720502.2007.10700478

M3 - Article

VL - 10

SP - 131

EP - 142

JO - Journal of Interdisciplinary Mathematics

JF - Journal of Interdisciplinary Mathematics

SN - 0972-0502

IS - 1

ER -