Sudoku sarežģītības noteikšanas algoritma izstrāde
Loading...
Date
Authors
Advisor
Journal Title
Journal ISSN
Volume Title
Publisher
Latvijas Universitāte
Language
lav
Abstract
Bakalaura darbā aprakstīta klasisku Sudoku mīklu sarežģītības noteicēja izstrāde un testēšana, salīdzinot tā sniegtos sarežģītības vērtējumus ar cilvēku sniegtajiem. Darba mērķis bija iegūt noteicēju, kas spēj atrisināt vismaz 99% no visām Sudoku mīklām un korekti noteikt sarežģītību vismaz 95% no mīklām. Darbā aprakstīti esoši Sudoku mīklu risināšanas un sarežģītības noteikšanas algoritmi un veikti secinājumi par to efektivitāti un ieteiktiem uzlabojumiem. Balstoties uz šiem secinājumiem izstrādāts jauns sarežģītības noteicējs, kas savu vērtējumu balsta uz risināšanas soļiem, kādus izmanto cilvēki, kad risina mīklas. Izstrādātais rīks tiek salīdzināts ar cilvēku sniegtiem Sudoku mīklu sarežģītības vērtējumiem.
This thesis contains the description of a difficulty determiner for classic Sudoku puzzles, as well as its creation and testing, comparing its given difficulty scores with those given by human solvers. The goal of this thesis was to create a determiner with the ability to solve at least 99% of all Sudoku puzzles and correctly determine difficulty for 95% of them. Multiple different Sudoku solving and difficulty determining algorithms have been described and conclusions have been made about their effectivity and proposed improvements. Based on these conclusions a new difficulty determiner has been created, based on the solving steps used by human solvers. The created tool has been compared to difficulty scores given by human Sudoku solvers.
This thesis contains the description of a difficulty determiner for classic Sudoku puzzles, as well as its creation and testing, comparing its given difficulty scores with those given by human solvers. The goal of this thesis was to create a determiner with the ability to solve at least 99% of all Sudoku puzzles and correctly determine difficulty for 95% of them. Multiple different Sudoku solving and difficulty determining algorithms have been described and conclusions have been made about their effectivity and proposed improvements. Based on these conclusions a new difficulty determiner has been created, based on the solving steps used by human solvers. The created tool has been compared to difficulty scores given by human Sudoku solvers.