An adaptive multi meme memetic algorithm embedding choice function hyper-heuristic

Heuristic methodologies appears for solving optimisation problems. Hyper-heuristics focus on search spaces to select or generate the suitable low-level heuristics to solve computationally difficult problems rather than focusing on finding solutions directly. The main goal is to develop more general...

Full description

Bibliographic Details
Main Author: Qarout, Rehab
Format: Dissertation (University of Nottingham only)
Language:English
Published: 2015
Subjects:
Online Access:https://eprints.nottingham.ac.uk/30793/
_version_ 1848794061320749056
author Qarout, Rehab
author_facet Qarout, Rehab
author_sort Qarout, Rehab
building Nottingham Research Data Repository
collection Online Access
description Heuristic methodologies appears for solving optimisation problems. Hyper-heuristics focus on search spaces to select or generate the suitable low-level heuristics to solve computationally difficult problems rather than focusing on finding solutions directly. The main goal is to develop more generally applicable search methodologies. The selection hyper-heuristics will be the core of designing the proposed algorithm and consist of two stages: the selection stage of low-level heuristics, and the acceptance stage of the solutions. An Evolutionary Algorithm approach produces high quality hyper-heuristics which can find optimal solutions for optimisation problems effectively. The Memetic Algorithms are evolutionary intelligent algorithms combining Genetic Algorithm with local search components. A Multi-Meme Memetic Algorithm presented in this project as a population based search method with Choice Function as a selection mechanism for low-level heuristics. The selection mechanism is encoded by multi-meme self-adaptation strategy for automating tuning of the choice function parameters. For each individual in the population, a meme encodes which setting is the best for Choice Function parameters for each operator type and relevant parameters of a chosen operator. Multi-Meme strategy is considered as a self-adaptive mechanism using a reward points system to increase the score for the meme that shows local improvement and uses these scores in the selection process. The proposed hyper-heuristics is tested and compared with the performance of previous hyper-heuristics which competed in the CHeSC2011 challenge across 9 problem domains. The achieved result was remarkable in some problem domains and opens some scope for further improvement in the proposed hyper-heuristic to improve the result in the rest of the problem domains.
first_indexed 2025-11-14T19:10:12Z
format Dissertation (University of Nottingham only)
id nottingham-30793
institution University of Nottingham Malaysia Campus
institution_category Local University
language English
last_indexed 2025-11-14T19:10:12Z
publishDate 2015
recordtype eprints
repository_type Digital Repository
spelling nottingham-307932017-10-19T15:05:49Z https://eprints.nottingham.ac.uk/30793/ An adaptive multi meme memetic algorithm embedding choice function hyper-heuristic Qarout, Rehab Heuristic methodologies appears for solving optimisation problems. Hyper-heuristics focus on search spaces to select or generate the suitable low-level heuristics to solve computationally difficult problems rather than focusing on finding solutions directly. The main goal is to develop more generally applicable search methodologies. The selection hyper-heuristics will be the core of designing the proposed algorithm and consist of two stages: the selection stage of low-level heuristics, and the acceptance stage of the solutions. An Evolutionary Algorithm approach produces high quality hyper-heuristics which can find optimal solutions for optimisation problems effectively. The Memetic Algorithms are evolutionary intelligent algorithms combining Genetic Algorithm with local search components. A Multi-Meme Memetic Algorithm presented in this project as a population based search method with Choice Function as a selection mechanism for low-level heuristics. The selection mechanism is encoded by multi-meme self-adaptation strategy for automating tuning of the choice function parameters. For each individual in the population, a meme encodes which setting is the best for Choice Function parameters for each operator type and relevant parameters of a chosen operator. Multi-Meme strategy is considered as a self-adaptive mechanism using a reward points system to increase the score for the meme that shows local improvement and uses these scores in the selection process. The proposed hyper-heuristics is tested and compared with the performance of previous hyper-heuristics which competed in the CHeSC2011 challenge across 9 problem domains. The achieved result was remarkable in some problem domains and opens some scope for further improvement in the proposed hyper-heuristic to improve the result in the rest of the problem domains. 2015-12-10 Dissertation (University of Nottingham only) NonPeerReviewed application/pdf en https://eprints.nottingham.ac.uk/30793/1/Rehab_Qarout_4201424.pdf Qarout, Rehab (2015) An adaptive multi meme memetic algorithm embedding choice function hyper-heuristic. [Dissertation (University of Nottingham only)] Hyper-heuristics Self-Adaptation Multi-Meme Memetic Algorithm Choice Function Heuristic Selection Cross-domain Optimisation.
spellingShingle Hyper-heuristics
Self-Adaptation
Multi-Meme
Memetic Algorithm
Choice Function
Heuristic Selection
Cross-domain Optimisation.
Qarout, Rehab
An adaptive multi meme memetic algorithm embedding choice function hyper-heuristic
title An adaptive multi meme memetic algorithm embedding choice function hyper-heuristic
title_full An adaptive multi meme memetic algorithm embedding choice function hyper-heuristic
title_fullStr An adaptive multi meme memetic algorithm embedding choice function hyper-heuristic
title_full_unstemmed An adaptive multi meme memetic algorithm embedding choice function hyper-heuristic
title_short An adaptive multi meme memetic algorithm embedding choice function hyper-heuristic
title_sort adaptive multi meme memetic algorithm embedding choice function hyper-heuristic
topic Hyper-heuristics
Self-Adaptation
Multi-Meme
Memetic Algorithm
Choice Function
Heuristic Selection
Cross-domain Optimisation.
url https://eprints.nottingham.ac.uk/30793/