研究生: |
孫立謙 Sun, Li-Chien |
---|---|
論文名稱: |
殺手數獨及其策略調查:朝向難度評級與人類友善的謎題設計 A Survey on Killer Sudoku and Its Strategies: Towards Difficulty Rating and Human-Friendly Puzzle Design |
指導教授: |
韓永楷
Hon, Wing-Kai |
口試委員: |
王弘倫
Wang, Hung-Lung 陳柏安 Chen, Po-An |
學位類別: |
碩士 Master |
系所名稱: |
電機資訊學院 - 資訊工程學系 Computer Science |
論文出版年: | 2024 |
畢業學年度: | 112 |
語文別: | 英文 |
論文頁數: | 37 |
中文關鍵詞: | 數獨 、殺手數獨 、解謎遊戲 、人類行為 、策略 、等級 、謎題設計 |
外文關鍵詞: | Sudoku, Killer Sudoku, Puzzle Game, Human-Behavior, Strategy, Rating, Puzzle Design |
相關次數: | 點閱:51 下載:0 |
分享至: |
查詢本校圖書館目錄 查詢臺灣博碩士論文知識加值系統 勘誤回報 |
這篇論文探討了殺手數獨的解題方法和策略,並提出了一個基於人類行為的解題模型。通過大量的調查網頁以及觀看其他玩家解題過程,本論文蒐集了適合人類使用的殺手數獨策略,另外,藉由觀察了數獨高手比賽時的行為、閱讀記憶力等相關研究,再搭配作者個人的遊戲經驗,本論文定義了適合人類的參數,以此建立了人類行為的解題模型。此模型通過定義一系列的邏輯步驟和決策點,提供了一個系統性解決殺手數獨謎題的流程框架。
本研究還探討了該模型在難度評分和謎題生成方面的應用。在評分規則中,我們可通過計算使用不同策略的次數和每個策略相應的分數來計算總分,用以評估謎題的難度。而在謎題生成方面,我們結合兩種常見數獨生成方法以作為殺手數獨生成的方法,以確保生成的謎題可以被人類解開並且具有唯一解。
This thesis explores methods and strategies for solving Killer Sudoku and proposes a human behavior solving model. We collected Killer Sudoku strategies suitable for humans through extensive research on various websites and observation of other players' solving processes. Additionally, we studied the behavior of Sudoku experts during competitions, and the research on human memory abilities, and then integrated these findings with the author's own gaming experience to establish parameters that are suitable for humans. This led to the development of a human behavior solving model, which offers a systematic framework for solving Killer Sudoku puzzles by outlining a series of logical steps and decision points.
This study also discusses the application of this model in difficulty rating and puzzle generation. We proposed a scoring system that calculates the total score by counting the frequency of different strategies used and their corresponding scores to examine the difficulty of puzzles. Furthermore, we proposed a method for generating Killer Sudoku puzzles by combining two common Sudoku generation methods to ensure that the generated puzzles are solvable by humans and have a unique solution.
[1] Bjorn van de Sand. Model for Human Killer Sudoku Solving. B.S. thesis, Utrecht University, 2018.
[2] David Eppstein. Nonrepetitive Paths and Cycles in Graphs with Application to Sudoku, 2005. arXiv:cs/0507053.
[3] Nelishia Pillay. Finding Solutions to Sudoku Puzzles using Human Intuitive Heuristics. South African Computer Journal, 49(1):25–34, 2012.
[4] Jose Barahona da Fonseca. From Mathematical Programming to Artificial Intelligence: A Novel MILP Model to Solve Killer Samurai Sudoku Puzzles. In Proceedings of International Conference on Advanced Communications and Computation, pages 12–17, 2016.
[5] Radek Pel´anek. Difficulty Rating of Sudoku Puzzles: An Overview and Evaluation, 2014. arXiv:1403.7373.
[6] Timo Mantere and Janne Koljonen. Solving, Rating and Generating Sudoku Puzzles with GA. In Proceedings of IEEE Congress on Evolutionary Computation, pages 1382–1389, 2007.
[7] Tan Tan Dai. Youtube Channel of Tan Tan Dai. https://www.youtube.com/@tantandai, 2018.
[8] Thomas Snyder. Youtube Channel of Thomas Snyder. https://www.youtube.com/@drsudoku, 2011.
[9] Nelson Cowan. The Magical Number 4 in Short-Term Memory: A Reconsideration of Mental Storage Capacity. Behavioral and Brain Sciences, 24(1):87–114, 2001.
[10] Nicolas Juillerat. Sukaku (Pencilmark Sudoku) and Sudoku Solver/Rater. https://github.com/SudokuMonster/SukakuExplainer, 2021.
[11] James Gibbs. Website of Killer Sudoku Puzzle. https://www.dailykillersudoku.com/, 2017.