Statistical determination of significant particle swarm optimization parameters: the case of Weibull distribution

Alptekin B., Acitas S., ŞENOĞLU B., ALADAĞ Ç. H.

SOFT COMPUTING, cilt.26, sa.22, ss.12623-12634, 2022 (SCI-Expanded) identifier identifier

  • Yayın Türü: Makale / Tam Makale
  • Cilt numarası: 26 Sayı: 22
  • Basım Tarihi: 2022
  • Doi Numarası: 10.1007/s00500-022-07253-y
  • Dergi Adı: SOFT COMPUTING
  • Derginin Tarandığı İndeksler: Science Citation Index Expanded (SCI-EXPANDED), Scopus, Academic Search Premier, Applied Science & Technology Source, Compendex, Computer & Applied Sciences, INSPEC, zbMATH
  • Sayfa Sayıları: ss.12623-12634
  • Anahtar Kelimeler: Particle swarm optimization, Parameter selection, Factorial design, Monte Carlo simulation, Weibull distribution, MODIFIED MAXIMUM-LIKELIHOOD, METAHEURISTIC ALGORITHM, SEARCH
  • Hacettepe Üniversitesi Adresli: Evet

Özet

Determination of the parameters of metaheuristic methods, such as particle swarm optimization (PSO) is a vital issue. Determining the optimal values of these parameters is a very difficult task and sometimes impossible so we focus on identifying the most important PSO parameters. For this purpose, statistical analysis of the relationship between the performance of the PSO algorithm with the proper selection of the PSO parameters should be done. Thus, unlike the earlier studies, here the most important PSO parameters are statistically determined. In the related literature, PSO parameters are traditionally determined by using trial and error method or specified intuitively. In this study, a new approach for identifying the most important PSO parameters for any optimization problem is proposed by using the factorial design which is one of the most popular and widely used statistical technique in experimental studies. Here, the parameters of the Weibull distribution which has a wide usage in the engineering and statistics literature are estimated just for illustration. It is obvious that identifying the most important PSO parameters via the proposed approach is very important to reduce the dimension of the search space for the PSO parameters. Statistical results obtained from the Monte Carlo simulation study are given and discussed at the end of the paper.