Hybrid bee colony optimization for examination timetabling problems

M. Alzaqebah, Salwani Abdullah

Research output: Contribution to journalArticle

28 Citations (Scopus)

Abstract

Swarm intelligence is a branch of artificial intelligence that focuses on the actions of agents in self-organized systems. Researchers have proposed a bee colony optimization (BCO) algorithm as part of swarm intelligence. BCO is a meta-heuristic algorithm based on the foraging behavior of bees. This study presents a hybrid BCO algorithm for examination timetabling problems. Bees in the BCO algorithm perform two main actions: forward pass and backward pass. Each bee explores the search space in forward pass and then shares information with other bees in the hive in backward pass. This study found that a bee decides to be either a recruiter that searches for a food source or a follower that selects a recruiter bee to follow on the basis of roulette wheel selection. In forward pass, BCO is supported along with other local searches, including the Late Acceptance Hill Climbing and Simulated Annealing algorithms. We introduce three selection strategies (tournament, rank and disruptive selection strategies) for the follower bees to select a recruiter to maintain population diversity in backward pass. The disruptive selection strategy outperforms tournament and rank selections. We also introduce a self-adaptive mechanism to select a neighborhood structure to enhance the neighborhood search. The proposed algorithm is evaluated against the latest methodologies in the literature with respect to two standard examination timetabling problems, namely, uncapacitated and competition datasets. We demonstrate that the proposed algorithm produces one new best result on uncapacitated datasets and comparable results on competition datasets.

Original languageEnglish
Pages (from-to)142-154
Number of pages13
JournalComputers and Operations Research
Volume54
DOIs
Publication statusPublished - 2015

Fingerprint

Timetabling
Optimization
Optimization Algorithm
Swarm Intelligence
Tournament
Roulette
Population Diversity
Neighborhood Search
Hill Climbing
Foraging
Simulated Annealing Algorithm
Metaheuristics
Heuristic algorithm
Wheel
Local Search
Search Space
Heuristic algorithms
Artificial Intelligence
Simulated annealing
Branch

Keywords

  • Bee colony optimization algorithm
  • Examination timetabling problems
  • Late acceptance hill climbing algorithm
  • Selection strategy
  • Self-adaptive mechanism
  • Simulated annealing

ASJC Scopus subject areas

  • Computer Science(all)
  • Modelling and Simulation
  • Management Science and Operations Research

Cite this

Hybrid bee colony optimization for examination timetabling problems. / Alzaqebah, M.; Abdullah, Salwani.

In: Computers and Operations Research, Vol. 54, 2015, p. 142-154.

Research output: Contribution to journalArticle

@article{eaff5ac1637d4282bc61e7872bf05431,
title = "Hybrid bee colony optimization for examination timetabling problems",
abstract = "Swarm intelligence is a branch of artificial intelligence that focuses on the actions of agents in self-organized systems. Researchers have proposed a bee colony optimization (BCO) algorithm as part of swarm intelligence. BCO is a meta-heuristic algorithm based on the foraging behavior of bees. This study presents a hybrid BCO algorithm for examination timetabling problems. Bees in the BCO algorithm perform two main actions: forward pass and backward pass. Each bee explores the search space in forward pass and then shares information with other bees in the hive in backward pass. This study found that a bee decides to be either a recruiter that searches for a food source or a follower that selects a recruiter bee to follow on the basis of roulette wheel selection. In forward pass, BCO is supported along with other local searches, including the Late Acceptance Hill Climbing and Simulated Annealing algorithms. We introduce three selection strategies (tournament, rank and disruptive selection strategies) for the follower bees to select a recruiter to maintain population diversity in backward pass. The disruptive selection strategy outperforms tournament and rank selections. We also introduce a self-adaptive mechanism to select a neighborhood structure to enhance the neighborhood search. The proposed algorithm is evaluated against the latest methodologies in the literature with respect to two standard examination timetabling problems, namely, uncapacitated and competition datasets. We demonstrate that the proposed algorithm produces one new best result on uncapacitated datasets and comparable results on competition datasets.",
keywords = "Bee colony optimization algorithm, Examination timetabling problems, Late acceptance hill climbing algorithm, Selection strategy, Self-adaptive mechanism, Simulated annealing",
author = "M. Alzaqebah and Salwani Abdullah",
year = "2015",
doi = "10.1016/j.cor.2014.09.005",
language = "English",
volume = "54",
pages = "142--154",
journal = "Surveys in Operations Research and Management Science",
issn = "0305-0548",
publisher = "Elsevier Limited",

}

TY - JOUR

T1 - Hybrid bee colony optimization for examination timetabling problems

AU - Alzaqebah, M.

AU - Abdullah, Salwani

PY - 2015

Y1 - 2015

N2 - Swarm intelligence is a branch of artificial intelligence that focuses on the actions of agents in self-organized systems. Researchers have proposed a bee colony optimization (BCO) algorithm as part of swarm intelligence. BCO is a meta-heuristic algorithm based on the foraging behavior of bees. This study presents a hybrid BCO algorithm for examination timetabling problems. Bees in the BCO algorithm perform two main actions: forward pass and backward pass. Each bee explores the search space in forward pass and then shares information with other bees in the hive in backward pass. This study found that a bee decides to be either a recruiter that searches for a food source or a follower that selects a recruiter bee to follow on the basis of roulette wheel selection. In forward pass, BCO is supported along with other local searches, including the Late Acceptance Hill Climbing and Simulated Annealing algorithms. We introduce three selection strategies (tournament, rank and disruptive selection strategies) for the follower bees to select a recruiter to maintain population diversity in backward pass. The disruptive selection strategy outperforms tournament and rank selections. We also introduce a self-adaptive mechanism to select a neighborhood structure to enhance the neighborhood search. The proposed algorithm is evaluated against the latest methodologies in the literature with respect to two standard examination timetabling problems, namely, uncapacitated and competition datasets. We demonstrate that the proposed algorithm produces one new best result on uncapacitated datasets and comparable results on competition datasets.

AB - Swarm intelligence is a branch of artificial intelligence that focuses on the actions of agents in self-organized systems. Researchers have proposed a bee colony optimization (BCO) algorithm as part of swarm intelligence. BCO is a meta-heuristic algorithm based on the foraging behavior of bees. This study presents a hybrid BCO algorithm for examination timetabling problems. Bees in the BCO algorithm perform two main actions: forward pass and backward pass. Each bee explores the search space in forward pass and then shares information with other bees in the hive in backward pass. This study found that a bee decides to be either a recruiter that searches for a food source or a follower that selects a recruiter bee to follow on the basis of roulette wheel selection. In forward pass, BCO is supported along with other local searches, including the Late Acceptance Hill Climbing and Simulated Annealing algorithms. We introduce three selection strategies (tournament, rank and disruptive selection strategies) for the follower bees to select a recruiter to maintain population diversity in backward pass. The disruptive selection strategy outperforms tournament and rank selections. We also introduce a self-adaptive mechanism to select a neighborhood structure to enhance the neighborhood search. The proposed algorithm is evaluated against the latest methodologies in the literature with respect to two standard examination timetabling problems, namely, uncapacitated and competition datasets. We demonstrate that the proposed algorithm produces one new best result on uncapacitated datasets and comparable results on competition datasets.

KW - Bee colony optimization algorithm

KW - Examination timetabling problems

KW - Late acceptance hill climbing algorithm

KW - Selection strategy

KW - Self-adaptive mechanism

KW - Simulated annealing

UR - http://www.scopus.com/inward/record.url?scp=85000197779&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85000197779&partnerID=8YFLogxK

U2 - 10.1016/j.cor.2014.09.005

DO - 10.1016/j.cor.2014.09.005

M3 - Article

AN - SCOPUS:85000197779

VL - 54

SP - 142

EP - 154

JO - Surveys in Operations Research and Management Science

JF - Surveys in Operations Research and Management Science

SN - 0305-0548

ER -