Improved orthogonal array based simulated annealing for design optimization

K. Y. Chan, Chun Kit Kwong, X. G. Luo

Research output: Journal article publicationJournal articleAcademic researchpeer-review

22 Citations (Scopus)


Recent research shows that simulated annealing with orthogonal array based neighbourhood functions can help in the search for a solution to a parametrical problem which is closer to an optimum when compared with conventional simulated annealing. Previous studies of simulated annealing analyzed only the main effects of variables of parametrical problems. In fact, both main effects of variables and interactions between variables should be considered, since interactions between variables exist in many parametrical problems. In this paper, an improved orthogonal array based neighbourhood function (IONF) for simulated annealing with the consideration of interaction effects between variables is described. After solving a set of parametrical benchmark function problems where interaction effects between variables exist, results of the benchmark tests show that the proposed simulated annealing algorithm with the IONF outperforms significantly both the simulated annealing algorithms with the existing orthogonal array based neighbourhood functions and the standard neighbourhood functions. Finally, the improved orthogonal array based simulated annealing was applied on the optimization of emulsified dynamite packing-machine design by which the applicability of the algorithm in real world problems can be evaluated and its effectiveness can be further validated.
Original languageEnglish
Pages (from-to)7379-7389
Number of pages11
JournalExpert Systems with Applications
Issue number4
Publication statusPublished - 1 May 2009


  • Design optimization
  • Interaction analysis
  • Neighbourhood functions
  • Orthogonal array
  • Simulated annealing

ASJC Scopus subject areas

  • Artificial Intelligence
  • Computer Science Applications
  • General Engineering


Dive into the research topics of 'Improved orthogonal array based simulated annealing for design optimization'. Together they form a unique fingerprint.

Cite this