Please use this identifier to cite or link to this item:
https://research.matf.bg.ac.rs/handle/123456789/3243| Title: | Automatsko rešavanje igre Sokoban pomoću veštačke inteligencije | Other Titles: | Automated solution of Sokoban puzzle using Artificial Intelligence | Authors: | Banković, Milan | Affiliations: | Informatics and Computer Science | Issue Date: | 2021 | Rank: | M63 | Publisher: | Beograd : Društvo za informacione sisteme i računarske mreže | Related Publication(s): | YU INFO 2021 : 27. Konferencija i izložba : Zbornik radova | Conference: | Konferencija YU INFO (27 ; 2021 ; Kopaonik) | Abstract: | Igra Sokoban je pomerajuća slagalica u kojoj igrač, čuvar skladišta, pomera kutije po tabli za igru, u cilju njihovog postavljanja na unapred zadata mesta. Pored toga što je zanimljiva za ručno rešavanje, ova igra inspiriše mnoge istraživače da se bave efikasnom automatizacijom njenog rešavanja, s obzirom na izuzetno veliku algoritamsku složenost ovog problema. U ovom radu, prikazujemo novi pristup zasnovan na automatskom planiranju uz pomoć procedura za ispitivanje zadovoljivosti u odnosu na teorije prvog reda, poznatih i kao SMT rešavači. The game of Sokoban is a sliding puzzle where a player, a warehouse keeper, pushes a number of boxes over the game board, in order to place them on the particular fields, given in advance. Besides the interest for manual solving, the game also attracts a great attention of the researchers to work on efficent automatization of its solving, because of the great algorithm complexity of the problem. In this paper, we propose a novel approach based on automated planning using procedures for checking satisfiability modulo first order theories, known as SMT solvers. |
URI: | https://research.matf.bg.ac.rs/handle/123456789/3243 |
| Appears in Collections: | Research outputs |
Show full item record
Google ScholarTM
Check
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.