Optimizacija problema raspoređivanja primjenom genetičkog algoritma s nišama.

doctoral thesis

• Author: Arijan Abrashi; VIAF: 305629592
• Mentor: Nedeljko Štefanić (mentor)
• Committee member: Nikola Šakić (predsjednik povjerenstva)
• Committee member: Nedeljko Štefanić (član povjerenstva)
• Committee member: Branko Novaković (član povjerenstva)
• Committee member: Ivo Čala (član povjerenstva)
• Committee member: Marin Golub (član povjerenstva)
• Granter: University of Zagreb; Faculty of Mechanical Engineering and Naval Architecture; Zagreb
• Defense date and country: 2011-05-02, Croatia
• Scientific / art field, discipline and subdiscipline: TECHNICAL SCIENCES; Mechanical Engineering
• Universal decimal classification (UDC): 658 - Business management, administration. Commercial organization

Abstract

U doktorskom su radu dana osnovna svojstva genetičkoga algoritma te njegove prednosti i nedostaci. Posebno su razmatrane metode selekcije, operatori križanja i mutacije te direktan i indirektan način prezentacije kromosoma. Također je analizirana razlika između haploidnoga i diploidnoga kromosoma te je dana analiza utjecaja diploidnoga kromosoma na očuvanje genetske raznolikosti populacije. Predložen je i ispitan genetički algoritam s nišama (GA) koji se za usporedbu jedinki u populaciji koristi takozvanom Hamiltonovom sličnošću. Prednost Hamiltonove sličnosti je u tome da ne postoji potreba za poznavanjem problema koji se rješava da bi se uspješno usporedila dva člana populacije (u konkretnom slučaju dva rasporeda). Algoritam je ispitan na dvije grupe poznatih benchmark problema, mt4 i la5. Dobiveni rezultati pokazuju manju standardnu devijaciju predloženog algoritma u odnosu prema jednostavnom genetičkom algoritmu (SGA), što jasno upozorava na njegovu superiornost. Osim Hamiltonove sličnosti, predloženo je i vremenski zavisno skaliranje funkcije cilja koje u suradnji s nišama znatno umanjuje mogućnost zapinjanja genetičkog algoritma u jednom od manje poželjnih lokalnih optimuma. Na kraju su date smjernice za daljnja istraživanja na području genetičkih algoritama s nišama.

Abstract (english)

In this PhD thesis basic characteristics of genetic algorithm have been presented along with its advantages and disadvantages. Selection methods, crossover and mutation operators, and direct and indirect chromosome presentation had especially been considered. The difference between haploid and diploid chromosomes was also analyzed and the influence of a diploid chromosome on the genetic diversity of a population presented. A niching genetic algorithm (GA), which uses so called Hamilton similarity for comparison of individuals in a population, was proposed and tested. The advantage of the Hamilton similarity lies in the fact that there is no need for context sensitive information in order to successfully compare two population members. Furthermore, the algorithm was tested on two famous benchmark problems (JSSP) - mt and la. The statistical results of the test were given. The significantly smaller standard deviation of the proposed GA compared to Simple GA clearly demonstrates its superiority. In addition to the Hamilton similarity, time dependent fitness scaling was also proposed, which in conjunction with niching significantly reduces the probability of the genetic algorithm getting trapped in one of the less desirable local optima. Finally, suggestions for future research are given.

• Keywords: genetički algoritam; niše; Hamiltonova sličnost; vremenski zavisno skaliranje funkcije cilja; raspoređivanje operacija po strojevima
• Keywords (english): genetic algorithm; niching; Hamilton similarity; time dependent fitness scaling; Job shop scheduling problem (JSSP)
• Language: croatian
• URN:NBN: urn:nbn:hr:235:650930
• Study programme: Title: Mechanical Engineering and Naval Architecture; Study programme type: university; Study level: postgraduate; Academic / professional title: doktor/doktorica znanosti, područje tehničkih znanosti (doktor/doktorica znanosti, područje tehničkih znanosti)
• Type of resource: Text
• File origin: Born digital
• Access conditions: Open access
• Created on: 2020-12-04 08:42:17