Pregled bibliografske jedinice broj: 1180209
A comparative study of dispatching rule representations in evolutionary algorithms
A comparative study of dispatching rule representations in evolutionary algorithms // IEEE Access (2022) doi:10.1109/access.2022.3151346 (znanstveni, prihvaćen)
CROSBI ID: 1180209 Za ispravke kontaktirajte CROSBI podršku putem web obrasca
Naslov
A comparative study of dispatching rule representations in evolutionary algorithms
Autori
Planinic, Lucija ; Backovic, Hrvoje ; Durasevic, Marko ; Jakobovic, Domagoj
Vrsta, podvrsta
Radovi u časopisima,
znanstveni
Izvornik
IEEE Access (2022)
Status rada
Prihvaćen
Ključne riječi
Unrelated machines environment ; Scheduling ; Solution representations ; Dispatching rules
Sažetak
Dispatching rules are most commonly used to solve scheduling problems under dynamic conditions. Since designing new dispatching rules is a time consuming process, it can be automatised by using various machine learning and evolutionary computation methods. In previous research, genetic programming was the most predominantly used method for the automatic design of new dispatching rules. However, there are many other evolutionary methods which use different representations than genetic programming, that can be used for generating dispatching rules. Some, like gene expression programming, were already successfully applied, while others like Cartesian genetic programming or grammatical evolution, were not used to generate dispatching rules. This paper will test six different methods (genetic programming, gene expression programming, Cartesian genetic programming, grammatical evolution, stack representation and analytic programming) to generate new dispatching rules for the unrelated machines environment, and will analyse how the tested methods perform on various scheduling criteria. The paper also analyses how different individual sizes in the tested methods affect the performance and average size of the generated dispatching rules. The results indicate that, except for the grammatical evolution and analytic programming, all tested methods perform quite similar, with the results depending on selected scheduling criterion. The results also demonstrate that Cartesian genetic programming was the most resistant to the occurrence of bloat, and that it evolved dispatching rules of the smallest average sizes.
Izvorni jezik
Engleski
Znanstvena područja
Računarstvo
POVEZANOST RADA
Projekti:
HRZZ-IP-2019-04-4333 - Hiperheurističko oblikovanje pravila raspoređivanja (HyDDRa) (Jakobović, Domagoj, HRZZ ) ( CroRIS)
Ustanove:
Fakultet elektrotehnike i računarstva, Zagreb
Citiraj ovu publikaciju:
Časopis indeksira:
- Current Contents Connect (CCC)
- Web of Science Core Collection (WoSCC)
- Science Citation Index Expanded (SCI-EXP)
- SCI-EXP, SSCI i/ili A&HCI
- Scopus