Combining particle swarm with ordinal optimization for stochastic simulation optimization problems

Shih Cheng Horng*, Feng-Yi Yang

*Corresponding author for this work

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

Abstract

In this paper, we combine the particle swarm (PS) with ordinal optimization (OO), abbreviated as CPSOO, to solve for a good enough solution of the stochastic simulation optimization problem (SSOP) with huge search space. First, a rough model using stochastic simulation with a small amount of test samples will be used as a fitness function evaluation in particle swarm optimization (PSO) algorithm to select N roughly good solutions from search space. Next, starting from the selected N roughly good solutions we proceed with goal softening procedure to search for a good enough solution. Finally, the proposed CPSOO algorithm is applied to a centralized broadband wireless network with k-limited service discipline, which is formulated as a SSOP that consists of a huge discrete search space comprised by the vector of k-limited service discipline. The vector of good enough k-limited service discipline obtained by the proposed algorithm is promising in the aspects of solution quality and computational efficiency.

Original languageEnglish
Title of host publicationASCC 2011 - 8th Asian Control Conference - Final Program and Proceedings
ChapterTuB2.1
Pages982-987
Number of pages6
StatePublished - 15 May 2011
Event8th Asian Control Conference, ASCC 2011 - Kaohsiung, Taiwan
Duration: 15 May 201118 May 2011

Publication series

NameASCC 2011 - 8th Asian Control Conference - Final Program and Proceedings

Conference

Conference8th Asian Control Conference, ASCC 2011
Country/TerritoryTaiwan
CityKaohsiung
Period15/05/1118/05/11

Keywords

  • centralized broadband wireless network
  • k-limited service discipline
  • ordinal optimization
  • particle swarm
  • stochastic simulation optimization

Fingerprint

Dive into the research topics of 'Combining particle swarm with ordinal optimization for stochastic simulation optimization problems'. Together they form a unique fingerprint.

Cite this