Multi-objective simulation optimization in a cluster/grid environment

Simulation optimization has received considerable attention due to the increased growth of manufacturing networks as well as global competition in the industry. Multi-objective Evolutionary Algorithms are developed to tackle these multi-objective optimization problems. Replication of simulation inst...

Full description

Saved in:
Bibliographic Details
Main Author: Tan, Wen Jun.
Other Authors: Stephen John Turner
Format: Final Year Project
Language:English
Published: 2010
Subjects:
Online Access:http://hdl.handle.net/10356/39798
Tags: Add Tag
No Tags, Be the first to tag this record!
Institution: Nanyang Technological University
Language: English
id sg-ntu-dr.10356-39798
record_format dspace
spelling sg-ntu-dr.10356-397982023-03-03T20:59:05Z Multi-objective simulation optimization in a cluster/grid environment Tan, Wen Jun. Stephen John Turner School of Computer Engineering Parallel and Distributed Computing Centre DRNTU::Engineering::Computer science and engineering::Computing methodologies::Simulation and modeling Simulation optimization has received considerable attention due to the increased growth of manufacturing networks as well as global competition in the industry. Multi-objective Evolutionary Algorithms are developed to tackle these multi-objective optimization problems. Replication of simulation instances are required so that the evolutionary algorithms can determine the optimal solutions with high confidence. The current process of determining the number of replications is only empirically selecting an arbitrarily value. Thus, the main goal of this project is to implement an algorithm to determine the number of replications required dynamically. A multi-objective computing budget allocation (MOCBA) procedure is chosen as it can determine the allocation according to the performance of the simulation. It is important to examine whether MOCBA sacrifices performance in order to obtain savings in computing budget. Therefore the performances of multi-objective evolutionary algorithms (MOEAs) integrated with the MOCBA procedure is compared against MOEAs integrated with Uniform Computing Budget Allocation (UCBA). Due to the time consuming computer simulations, mathematical test problems are used instead. A varying amount of noise is applied at the objective functions to test the performance and adaptability of the different evolutionary algorithms and computing budget allocation. After studying the algorithm with theoretical test problems, a case study on reverse logistics is used for evaluation. The importance of the reverse logistics network is often neglected by the companies in denial of the cost involved. An effective design of a closed-loop forward supply chain and reverse maintenance network can help to reduce cost and bring more value to the products. Therefore using the simulation model of this network is an effective way to illustrate the performance of the integrated evolutionary algorithm with MOCBA. Bachelor of Engineering (Computer Engineering) 2010-06-04T04:21:57Z 2010-06-04T04:21:57Z 2010 2010 Final Year Project (FYP) http://hdl.handle.net/10356/39798 en Nanyang Technological University 63 p. application/pdf
institution Nanyang Technological University
building NTU Library
continent Asia
country Singapore
Singapore
content_provider NTU Library
collection DR-NTU
language English
topic DRNTU::Engineering::Computer science and engineering::Computing methodologies::Simulation and modeling
spellingShingle DRNTU::Engineering::Computer science and engineering::Computing methodologies::Simulation and modeling
Tan, Wen Jun.
Multi-objective simulation optimization in a cluster/grid environment
description Simulation optimization has received considerable attention due to the increased growth of manufacturing networks as well as global competition in the industry. Multi-objective Evolutionary Algorithms are developed to tackle these multi-objective optimization problems. Replication of simulation instances are required so that the evolutionary algorithms can determine the optimal solutions with high confidence. The current process of determining the number of replications is only empirically selecting an arbitrarily value. Thus, the main goal of this project is to implement an algorithm to determine the number of replications required dynamically. A multi-objective computing budget allocation (MOCBA) procedure is chosen as it can determine the allocation according to the performance of the simulation. It is important to examine whether MOCBA sacrifices performance in order to obtain savings in computing budget. Therefore the performances of multi-objective evolutionary algorithms (MOEAs) integrated with the MOCBA procedure is compared against MOEAs integrated with Uniform Computing Budget Allocation (UCBA). Due to the time consuming computer simulations, mathematical test problems are used instead. A varying amount of noise is applied at the objective functions to test the performance and adaptability of the different evolutionary algorithms and computing budget allocation. After studying the algorithm with theoretical test problems, a case study on reverse logistics is used for evaluation. The importance of the reverse logistics network is often neglected by the companies in denial of the cost involved. An effective design of a closed-loop forward supply chain and reverse maintenance network can help to reduce cost and bring more value to the products. Therefore using the simulation model of this network is an effective way to illustrate the performance of the integrated evolutionary algorithm with MOCBA.
author2 Stephen John Turner
author_facet Stephen John Turner
Tan, Wen Jun.
format Final Year Project
author Tan, Wen Jun.
author_sort Tan, Wen Jun.
title Multi-objective simulation optimization in a cluster/grid environment
title_short Multi-objective simulation optimization in a cluster/grid environment
title_full Multi-objective simulation optimization in a cluster/grid environment
title_fullStr Multi-objective simulation optimization in a cluster/grid environment
title_full_unstemmed Multi-objective simulation optimization in a cluster/grid environment
title_sort multi-objective simulation optimization in a cluster/grid environment
publishDate 2010
url http://hdl.handle.net/10356/39798
_version_ 1759854871388880896