Solving two-dimensional layout optimization problems with irregular shapes by using meta-heuristic
Solving two-dimensional layout optimization problems with irregular shapes by using meta-heuristic
The focus of this thesis is developing methodologies for 2-dimensional problems that involve irregular shapes, where the objective is to find an arrangement of the irregular pieces in order to minimise waste material. There are two main approaches popular with researchers; in this research project we will group them into iterative constructive heuristics (ICH) and those heuristics which search over the physical layout (SOL). ICH seeks to generate good layouts by placing pieces on the stock sheet piece by piece according to a placement rule. The orders the pieces are placed is controlled by a search algorithm. SOL works with the physical layout and tries to improve the solution by moving the placement position of pieces. Overlap is often permitted and penalised in the cost function in order to generate new solutions. Both approaches are competitive and each new publication brings better results with respect to the benchmark data sets. Although this can be credited to better algorithm design, it could also be argued that researchers are getting better at incorporating sophisticated specific features in their algorithms to handle the 17 benchmark data sets. In this research we intend to investigate the two representations of the problem and establish some principles of the strengths and weaknesses of each method with respect to data type. In order to conduct this research the algorithms will be developed using only the basic principles of both approaches and discarding any special features found in the literature. The aim is to deduct from the experimental results an understanding of what solution representation should be applied given the data type, performance requirements and number of pieces.
University of Southampton
Ramakrishnan, Kumaran
16f01b4e-0bc5-4f1f-8002-7ca367952474
2008
Ramakrishnan, Kumaran
16f01b4e-0bc5-4f1f-8002-7ca367952474
Ramakrishnan, Kumaran
(2008)
Solving two-dimensional layout optimization problems with irregular shapes by using meta-heuristic.
University of Southampton, Doctoral Thesis.
Record type:
Thesis
(Doctoral)
Abstract
The focus of this thesis is developing methodologies for 2-dimensional problems that involve irregular shapes, where the objective is to find an arrangement of the irregular pieces in order to minimise waste material. There are two main approaches popular with researchers; in this research project we will group them into iterative constructive heuristics (ICH) and those heuristics which search over the physical layout (SOL). ICH seeks to generate good layouts by placing pieces on the stock sheet piece by piece according to a placement rule. The orders the pieces are placed is controlled by a search algorithm. SOL works with the physical layout and tries to improve the solution by moving the placement position of pieces. Overlap is often permitted and penalised in the cost function in order to generate new solutions. Both approaches are competitive and each new publication brings better results with respect to the benchmark data sets. Although this can be credited to better algorithm design, it could also be argued that researchers are getting better at incorporating sophisticated specific features in their algorithms to handle the 17 benchmark data sets. In this research we intend to investigate the two representations of the problem and establish some principles of the strengths and weaknesses of each method with respect to data type. In order to conduct this research the algorithms will be developed using only the basic principles of both approaches and discarding any special features found in the literature. The aim is to deduct from the experimental results an understanding of what solution representation should be applied given the data type, performance requirements and number of pieces.
Text
1282382.pdf
- Version of Record
More information
Published date: 2008
Identifiers
Local EPrints ID: 466662
URI: http://eprints.soton.ac.uk/id/eprint/466662
PURE UUID: 40c21d9a-3dfe-4d3f-987d-2efc506ed8a4
Catalogue record
Date deposited: 05 Jul 2022 06:16
Last modified: 16 Mar 2024 20:50
Export record
Contributors
Author:
Kumaran Ramakrishnan
Download statistics
Downloads from ePrints over the past year. Other digital versions may also be available to download e.g. from the publisher's website.
View more statistics