dc.contributor.authorHuang, Youliangen_US
dc.identifier.citationHuang, Y. (2007). Machine learning in customer-driven product configuration based on lifecycle metrics. Master’s thesis, Nanyang Technological University, Singapore.
dc.description.abstractMass customization has become a major trend for the manufacturing sector today, as it enables companies to produce customized products that meet customers' individual requirements at mass production rate. Product configuration tool is an integral part of the manufacturing service system that has been recognized as the key enabler to achieve mass customization. Although intensively studied in the literature, the application of the product configuration is still in the initial stage, due to the lack of systematic modeling and efficient configuration approach. Heterogeneous product information in configuration has to be effectively managed. In this thesis, firstly we give an intensive study on the state-of-the-art computer-aided product configuration methodologies. The advantages and limitations of existing approaches are evaluated. Constraint-based approach is proposed to model and solve product configuration problem. It is based on Constraint Satisfaction Problem (CSP) paradigm, where problem is modeled into variables, domains and constraints. Based on the approach, a systematic product configuration system framework is then defined. Novel product data model and knowledge representation model are designed and developed to capture the domain problem. Moreover, to address the limitation of the manual knowledge acquisition approach for knowledge-based configuration system, a novel machine learning approach which deploys association rule mining technique is proposed and implemented to achieve automatic product configuration knowledge generation. A product configuration software system is developed to implement the constraint-based configuration approach proposed, which demonstrates the significance of our modeling and reasoning approach. Constraint-based configuration is proven to be an effective and efficient approach. To further enhance our approach, the system is extended to take cost as design criteria. A cost model that provides cost assessment for whole product life cyle has been developed. A system module is developed and integrated with the configuration system to provide cost estimation capability, which is a significant value-add for practical product configuration application.en_US
dc.format.extent102 p.en_US
dc.subjectDRNTU::Engineering::Computer science and engineering::Computer applications::Physical sciences and engineering
dc.titleMachine learning in customer-driven product configuration based on lifecycle metricsen_US
dc.contributor.schoolSchool of Computer Engineeringen_US
dc.contributor.supervisorNg Wee Keongen_US
dc.description.degreeMASTER OF ENGINEERING (SCE)en_US

Files in this item


This item appears in the following Collection(s)

Show simple item record