近年來，數位遊戲軟體蓬勃發展，大部份數位遊戲軟體都具有一些人工智慧的應用。由於遊戲人工智慧開發困難且複雜度高，長久以來應用於遊戲的人工智慧多較簡易難度有限，但是無法自我更新、成長的人工智慧只能被侷限在某種程度，使得遊戲開發者總是不斷增加難度或敵人數量來滿足玩家的需求。
本論文研究建置一個應用於戰略遊戲的人工智慧，以類神經網路為基礎計算戰場中的影響範圍，利用學習回饋的特性訓練電腦分析戰場上的影響範圍，使電腦在戰略遊戲中進行戰術模擬，根據分析後的結果去指揮士兵進行遊戲，讓電腦指揮官有類似人類玩家的學習行為，如此一來遊戲內容更具挑戰性。

In recent years, digital game software grows flourish, most of digital games software has designed artificial intelligence in game. Because the development of game artificial intelligence is difficult and complex, it is limited the growth of the game artificial intelligence. Most of artificial intelligence of digital game is without self-learning, so the game developers always increase the number of enemies to make the game difficult to be completed in short time.
In this thesis, we proposed an artificial intelligence of strategy simulation for war game. We use artificial neural networks to construct influence map, and train the computer to analysis the influence map. According to the result of a battle game, the system will feedback to the game artificial intelligence and learn the behavior of player. In this way, the game would become more challenging for user to play.

第一章 緒論	1
1.1	研究動機及研究目的	1
1.2	論文架構	4
第二章 相關研究	5
2.1	Reinforcement Learning	5
2.2	Genetic Algorithm	7
2.3	Bayesian Networks	9
2.4	Artificial Neural Network	12
第三章 研究內容與方法	15
3.1	系統架構	15
3.2	分析影響範圍	17
3.3	戰術評分	23
3.4	戰術權重	28
第四章 模擬實作與結果討論	32
4.1	模擬實作說明	32
4.1.1	模擬實作目標	32
4.1.2	模擬實作環境	33
4.1.3	模擬實作介面	34
4.2	模擬結果與討論	35
第五章 結論與未來相關研究	42
5.1	結論	42
5.2	未來相關研究	43
參考文獻	44
附錄—英文論文	46

圖目錄
圖2.1 The standard reinforcement-learning model	6
圖2.2 Genetic Algorithm 演化流程圖	8
圖2.3 Dynamic Difficulty Controlling Game System	9
圖2.4 Bayesian Networks結構	11
圖2.5 分析威脅偵測	12
圖2.6 生物神經細胞	13
圖2.7 類神經網路節點	13
圖2.8 類神經網路模型結構	14
圖3.1 系統架構圖	15
圖3.2 系統流程圖	16
圖3.3 士兵攻擊距離與移動範圍	18
圖3.4 基礎士兵影響範圍	18
圖3.5 遠距離士兵影響範圍	19
圖3.6 (a) 行軍狀態戰術拓樸	25
圖3.6 (b) 戰鬥狀態戰術拓樸	25
圖3.7 父節點Y的戰術價值結構	28
圖4.1 實做介面	34
圖4.2雙方陣營的影響範圍	34
圖4.3 無地形影響模擬畫面	35
圖4.4 無地形影響模擬過程畫面	35
圖4.5(a) 無地形影響權重權重變化折線圖(一)	36
圖4.5(b) 無地形影響權重權重變化折線圖(二)	37
圖4.6 無地形影響權重權重變化折線圖(三)	37
圖4.7無地形影響權重權重變化折線圖(四)	39
圖4.8一般地形模擬畫面	39
圖4.9一般地形模擬過程	40
圖4.10一般地形權重變化折線圖	41
圖4.11 SSAI一般地形模擬結果	41

表目錄
表 3.1 IM的資料結構	21
表 3.2 Influence-Move	22
表 3.3 Influence-Attack	22
表3.4戰術評分參數表	26
表4.1 環境平台軟硬體設備	33

參考文獻
References
中文部份：
[周榮 98] 周榮(1998)，電腦角色扮演遊戲中的神話原型之研究，國立交通大
學傳播研究所碩士論文。
英文部分：
[Bourg 04] David M. Bourg and Glenn Seemann, ”AI for Game Developers”,
O'Reilly Media, July 2004.
[Chao 06] Hsiao-Chuan Chao, ”A reinforcement self-learning model
on an intelligent behavior avatar in a virtual world”, Sensor Networks,
Ubiquitous, and Trustworthy Computing, 2006. IEEE International
Conference on 2006.
[Guardian 07] Guardian URL：The hard-thought race for intelligent gaming
http://www.guardian.co.uk/technology/2007/apr/19/games.guardianwe
eklytechnologysection3
[Hunt 00] Hunt, M.; von Konsky, B.; Venkatesh, S.; Petros, P., "Bayesian
networks and decision trees in the diagnosis of female urinary
incontinence", Engineering in Medicine and Biology Society, the 22nd
Annual International Conference of the IEEE, vol.1, 2000. pp551 –
554.
[Hung 03] Haw-Yun Hung, ”Building an intelligent behavior avater in a virtual
world”, Master thesis, University of Tamkang, Tansui, Taiwan, 2003.
[John 75] John Holland, ”Adaption in natural and artificial systems”, The
University Michigan Press, Ann Arbor. 1975.
[Jang 08] Su-Hyung Jang, Student Member, IEEE and Sung-Bae Cho, Senior
Member, IEEE,” Evolving Neural NPCs with Layered Influence Map
in the Real-time Simulation Game 'Conqueror'”, Computational
Intelligence and Games, 2008. pp. 385 – 388.
[Lin 00] Sinn-Cheng Lin, “ On Soft Computing Theory and Knowledge
Processing Technology,” Education Media and Library Sciences, Vol.
38, No. 2, Dec. 2000. pp. 148-173.
[Li 09] Li Man; Feng Xinxi, “Chen Jiaoping,Research of Threat Identification
Based on Bayesian Networks”, Wireless Communications,
Networking and Mobile Computing, WiCom '09. 5th, 2009. pp 1-3.
[Moore 96] Andrew W.Moore,Leslie Pack Kaelbling, and Michael L.Littman.,
“Reinforcement learning:A survey.”, Journal of Artificial Intelligence
Research, Vol. 4, 1996. pp. 237-285.
[Um 07] Sang-Won Um, Tae-Yong Kim, and Jong-Soo Choi,”Dynamic
Difficulty Controlling Game System”, IEEE Transactions on
Consumer Electronics, Vol. 53, No. 2, MAY 2007. pp. 812-818.
[Yang 04] Chia-Che Yang, ”Research and Development on Searching a Routing
Path of a Dynamic Terrain”, Master thesis, University of Tamkang,
Tansui, Taiwan, 2004.