Solving nurikabe game with ant colony optimization

V nalogi predstavljamo učinkovitost optimizacije s kolonijami mravelj za reševanje japonske miselne igre Nurikabe. Razložimo pravila igre in njene osnovne lastnosti ter optimizacijo s kolonijami mravelj. Pojasnimo lastno implementacijo metode in metode iskanja v globino. Analiziramo krmilne parametre optimizacije s kolonijami mravelj. Vse algoritme primerjamo med seboj. Iz dobljenih rezultatov smo ugotovili, da je iskanje v globino časovno zahtevno. Optimizacija s kolonijami mravelj dosega dobre rezultate in omogoča reševanje zahtevnejših problemov.In this thesis, the efficiency of the ant colony optimization method for solving the Japanese logic puzzle Nurikabe is presented. First, the rules and fundamental properties of the puzzle and the ant colony optimization technique are explained. Our implementation of the method as well as our depth-first search method are described. The input parameters of the algorithm are analyzed. The results of all the algorithms are compared. The depth-first search method is time consuming. Ant colony optimization brings better results and makes it possible to solve more difficult problems