Skip to main navigation Skip to search Skip to main content

Optimal wet-furnace machine allocation for daily fab production

  • Simon Wang
  • , Fenix Wang
  • , Joey Chang
  • , Chang Ju-Yin Chang
  • , Chang Shi-Chung Chang
  • , Wang Peng Wang
  • , Peter B. Luh
  • , Kao Yu-Ting Kao
  • , Zhan Shun-Cheng Zhan

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

3 Scopus citations

Abstract

This paper presents the design of an efficient and near-optimal machine allocation method for wet bench and furnace machine groups to achieve daily production targets and high moves. The core of the method consists of an innovative integer linear programming problem formulation of the complex operation constraints such as variable size batching, queue time, cross fab transportation and linearization of the objective function, and the skillful application of an optimization tool suite. Test results over real fab data clearly demonstrated the superiority of the method than a currently practiced heuristic in bottleneck (furnace) machine utilization (potentially by 10-15%) and it leads to nearly 50% reduction in average total waiting time per FOUP. Its computation efficiency (in seconds) indicates a strong potential for both solving larger sized problems and extensions to dynamic allocations.

Original languageEnglish
Title of host publication2010 International Symposium on Semiconductor Manufacturing, ISSM 2010
StatePublished - 2010
Externally publishedYes
Event18th International Symposium on Semiconductor Manufacturing, ISSM 2010 - Tokyo, Japan
Duration: 18 Oct 201020 Oct 2010

Publication series

NameIEEE International Symposium on Semiconductor Manufacturing Conference Proceedings
ISSN (Print)1523-553X

Conference

Conference18th International Symposium on Semiconductor Manufacturing, ISSM 2010
Country/TerritoryJapan
CityTokyo
Period18/10/1020/10/10

Fingerprint

Dive into the research topics of 'Optimal wet-furnace machine allocation for daily fab production'. Together they form a unique fingerprint.

Cite this