Topic-specific Web robot model based on restless bandits

T. O'Meara, A. Patel

Research output: Contribution to journalArticle

9 Citations (Scopus)

Abstract

Reinforcement-learning algorithm and restless-bandits generalization are used to construct and maintain topic-specific web indexes. The topic-specific we robot cshedules the documents requests to maximize the quality of its topic-specific document databases. The basic components of web robot are request controller and fetcher controller. Restless bandits is a generalization of a classical project-scheduling problem known as the multi-armed bandit problem. The performance of different algorithms was evaluated and fringe greedy algorithm's performance was also calulated. Results showed that the proposed algorithm is good to build the complete solution.

Original languageEnglish
Pages (from-to)27-35
Number of pages9
JournalIEEE Internet Computing
Volume5
Issue number2
DOIs
Publication statusPublished - Mar 2001
Externally publishedYes

Fingerprint

Software agents
Controllers
Reinforcement learning
Learning algorithms
Scheduling
Robots

ASJC Scopus subject areas

  • Computer Graphics and Computer-Aided Design
  • Software

Cite this

Topic-specific Web robot model based on restless bandits. / O'Meara, T.; Patel, A.

In: IEEE Internet Computing, Vol. 5, No. 2, 03.2001, p. 27-35.

Research output: Contribution to journalArticle

O'Meara, T. ; Patel, A. / Topic-specific Web robot model based on restless bandits. In: IEEE Internet Computing. 2001 ; Vol. 5, No. 2. pp. 27-35.
@article{1ab8cf714f954dd0ab51ce9339fc650d,
title = "Topic-specific Web robot model based on restless bandits",
abstract = "Reinforcement-learning algorithm and restless-bandits generalization are used to construct and maintain topic-specific web indexes. The topic-specific we robot cshedules the documents requests to maximize the quality of its topic-specific document databases. The basic components of web robot are request controller and fetcher controller. Restless bandits is a generalization of a classical project-scheduling problem known as the multi-armed bandit problem. The performance of different algorithms was evaluated and fringe greedy algorithm's performance was also calulated. Results showed that the proposed algorithm is good to build the complete solution.",
author = "T. O'Meara and A. Patel",
year = "2001",
month = "3",
doi = "10.1109/4236.914645",
language = "English",
volume = "5",
pages = "27--35",
journal = "IEEE Internet Computing",
issn = "1089-7801",
publisher = "Institute of Electrical and Electronics Engineers Inc.",
number = "2",

}

TY - JOUR

T1 - Topic-specific Web robot model based on restless bandits

AU - O'Meara, T.

AU - Patel, A.

PY - 2001/3

Y1 - 2001/3

N2 - Reinforcement-learning algorithm and restless-bandits generalization are used to construct and maintain topic-specific web indexes. The topic-specific we robot cshedules the documents requests to maximize the quality of its topic-specific document databases. The basic components of web robot are request controller and fetcher controller. Restless bandits is a generalization of a classical project-scheduling problem known as the multi-armed bandit problem. The performance of different algorithms was evaluated and fringe greedy algorithm's performance was also calulated. Results showed that the proposed algorithm is good to build the complete solution.

AB - Reinforcement-learning algorithm and restless-bandits generalization are used to construct and maintain topic-specific web indexes. The topic-specific we robot cshedules the documents requests to maximize the quality of its topic-specific document databases. The basic components of web robot are request controller and fetcher controller. Restless bandits is a generalization of a classical project-scheduling problem known as the multi-armed bandit problem. The performance of different algorithms was evaluated and fringe greedy algorithm's performance was also calulated. Results showed that the proposed algorithm is good to build the complete solution.

UR - http://www.scopus.com/inward/record.url?scp=0035271555&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0035271555&partnerID=8YFLogxK

U2 - 10.1109/4236.914645

DO - 10.1109/4236.914645

M3 - Article

AN - SCOPUS:0035271555

VL - 5

SP - 27

EP - 35

JO - IEEE Internet Computing

JF - IEEE Internet Computing

SN - 1089-7801

IS - 2

ER -