A FNP approach for evaluating and enhancing the long-term competitiveness of a semiconductor fabrication factory through yield learning modeling

Tin-Chih Chen*

*Corresponding author for this work

Research output: Contribution to journalArticle

13 Scopus citations

Abstract

A systematic procedure is established in this study to evaluate and enhance the long-term competitiveness of a semiconductor manufacturing factory. At first, after assessing the competitiveness of all its products at several check points, the long-term competitiveness of the semiconductor manufacturing factory can be derived. Subsequently, to further enhance the long-term competitiveness of the semiconductor manufacturing factory, a capacity re-allocation mechanism is designed to improve the competitiveness of less competitive products at the least expense of the other products that are highly competitive with dynamic capacity re-allocation. For this purpose, a fuzzy, nonlinear programming (FNP) model is constructed, which is then converted into an equivalent NP problem solved with existing optimization software. To evaluate the effectiveness of the proposed methodology, some data in a real semiconductor fabrication factory were collected. Experimental results revealed that the long-term competitiveness of the semiconductor fabrication factory could be evaluated by 2.4% if the factory capacity was re-allocated accordingly. Further, the established dynamic capacity re-allocation plan was shown to be very efficient.

Original languageEnglish
Pages (from-to)993-1003
Number of pages11
JournalInternational Journal of Advanced Manufacturing Technology
Volume40
Issue number9-10
DOIs
StatePublished - 1 Feb 2009

Keywords

  • Capacity re-allocation
  • Fuzzy nonlinear programming
  • Long-term competitiveness
  • Semiconductor
  • Yield learning

Fingerprint Dive into the research topics of 'A FNP approach for evaluating and enhancing the long-term competitiveness of a semiconductor fabrication factory through yield learning modeling'. Together they form a unique fingerprint.

Cite this