Optimization of scheduling for unrelated machines

U ovom radu opisan je problem raspoređivanja u okolini nesrodnih strojeva. Opisan je algoritam genetskog programiranja kao i raspoređvanje zasnovano na prilagodljivim pravilima koje zapravo predstavlja primjenu genetskog programiranja na izradu rasporeda. Opisan je niz optimizacija (optimizacija parametara, dodavanje novih funkcijskih čvorova, semantičko genetsko programiranje, iterativna pravila raspoređivanje i GEP) koje su implementirane i isprobane kako bi se postigla kvalitetnije izrada rasporeda od strane genetskog programiranja. Prikazani su i komentirani rezultati koji su dobiveni za svaku pojedinu optimizacijsku tehniku.In this thesis the problem of scheduling in the unrelated machines environment was described. The genetic programming algorithm as well as scheduling based on adaptive rules, which in itself is nothing but the application of genetic programming to generating schedules, were also described. A number of optimization techniques (parameter optimization, adding new functional nodes, semantic genetic programming, iterative dispatching rules and gene expression programming) which were implemented and tested in order to achieve improvements in generating the schedules, were also described. Finally, for each of the mentioned optimization techniques the results of it's usage are shown and commented upon