跳到主要內容

臺灣博碩士論文加值系統

(44.192.94.177) 您好!臺灣時間:2024/07/21 18:35
字體大小: 字級放大   字級縮小   預設字形  
回查詢結果 :::

詳目顯示

我願授權國圖
: 
twitterline
研究生:林子芃
研究生(外文):Zi-Pong Lim
論文名稱:深度強化學習在多人連線戰鬥遊戲中隊伍選角推薦的運用
論文名稱(外文):Deep Reinforcement Learning for Progressive Team Draft Recommendations in MOBA Games
指導教授:林守德林守德引用關係
指導教授(外文):Shou-De Lin
口試委員:解巽評駱宏毅李政德
口試委員(外文):Hsun-Ping HsiehHung-Yi LoCheng-Te Li
口試日期:2018-07-23
學位類別:碩士
校院名稱:國立臺灣大學
系所名稱:資訊工程學研究所
學門:工程學門
學類:電資工程學類
論文種類:學術論文
論文出版年:2018
畢業學年度:106
語文別:英文
論文頁數:23
中文關鍵詞:DOTAMOBA角色推薦
相關次數:
  • 被引用被引用:0
  • 點閱點閱:232
  • 評分評分:
  • 下載下載:0
  • 收藏至我的研究室書目清單書目收藏:0
MOBA是一種把玩家分成兩隊以互相進行對抗的電競遊戲種類。遊戲開始前,兩隊的隊員必須輪流從眾多的角色中選擇一個來操控。選角對MOBA來說是一個非常重要的環節,因為好的角色組合可以通過加成相互提升能力,而可以克制對手角色的角色則能讓對手能力大打折扣。在這篇論文裡,我們提出一個漸進式的角色推薦系統,可以在選角環節中的任何一個階段,針對目前隊友和對手已經選擇的角色,進行角色選擇推薦。業餘玩家和職業玩家都可以利用這套角色推薦系統在賽前做預測和制定戰略。遊戲開發者也可以利用這套系統對遊戲進行調整。這篇論文裡所有實驗都將會被進行在DOTA2,一個很流行的MOBA上。
Multiplayer Online Battle Arena (MOBA) is a genre of games in which two teams of players compete against each other. Both teams taking turns picking one draft at a time from a pool of characters before a match begins. The drafting phase can be a crucial phase, as a good combination of team characters and counter-combinations against the rival team, may highly improve the winning chances of a team. In this paper, we propose a progressive team draft recommendation system, that recommends one character to draft for a team, in each round of the drafting phase, given current enemy drafts and ally drafts. The recommendation system can be used by teams, both casual and professional, to make predictions and develop strategies before a match, or by game developers, to balance and fine-tune the game. All the experiments and results in this paper will be based on a popular MOBA called DOTA2.
1 Introduction 1
1.1 Background................................. 1
1.2 Game Mechanics .............................. 2
1.3 Research Motivation ............................ 3
1.4 Problem Statement ............................. 4
1.5 Overview .................................. 5
2 Related Works 6
2.1 Match Outcome Prediction ......................... 6
2.2 Team Draft Recommendation........................ 7
3 Methodology & Implementation 8
3.1 Dataset ................................... 8
3.2 Outcome Prediction Model......................... 9
3.3 Recommendation System.......................... 10
3.3.1 Reinforcement Learning ...................... 11
3.3.2 Problem Formulation to Reinforcement Learning . . . . . . . . . 13
3.3.3 Deep Reinforcement Learning................... 14
3.3.4 Tweaks ............................... 16
4 Evaluation
4.1 Experiment Setup.............................. 17
4.1.1 Recommendation Systems ..................... 17
4.1.2 Drafting Orders........................... 18
4.1.3 Enemy Picking Methods ...................... 18
4.1.4 Different Prediction Models .................... 19
4.2 Results.................................... 19
5 Conclusion 21
5.1 Summary .................................. 21
5.2 Future Work................................. 21
Bibliography 23
[1] Kevin E. Conley and Daniel Perry. How does he saw me ? a recommendation engine for picking heroes in dota 2. 2013.
[2] Kaushik Kalyanaraman. To win or not to win ? a prediction model to determine the outcome of a dota 2 match. 2015.
[3] Nicholas Kinkade. Dota 2 win prediction. 2015.
[4] Aleksandr Semenov, Peter Romov, Sergey Korolev, Daniil Yashkov, and Kirill Neklyudov. Performance of machine learning algorithms in predicting game out- come from drafts in dota 2. In Dmitry I. Ignatov, Mikhail Yu. Khachay, Valeri G. Labunets, Natalia Loukachevitch, Sergey I. Nikolenko, Alexander Panchenko, An- drey V. Savchenko, and Konstantin Vorontsov, editors, Analysis of Images, Social Networks and Texts, pages 26–37, Cham, 2017. Springer International Publishing.
[5] LucasHankeandLuizChaimowicz.Arecommendersystemforheroline-upsinmoba games. 2017.
QRCODE
 
 
 
 
 
                                                                                                                                                                                                                                                                                                                                                                                                               
第一頁 上一頁 下一頁 最後一頁 top