博碩士論文 111423017 詳細資訊




以作者查詢圖書館館藏 以作者查詢臺灣博碩士 以作者查詢全國書目 勘誤回報 、線上人數:145 、訪客IP:18.224.73.226
姓名 王駿豪(Chun-Hao Wang)  查詢紙本館藏   畢業系所 資訊管理學系
論文名稱 利用納許Q學習分析傳統廣告贊助與下一代付費服務之競爭:以Google與BingChat為例
相關論文
★ 直播平台之廣告競爭策略與利益分享制度分析★ 以雙邊平台模型分析雲端遊戲主機經營策略
★ 利用納許Q學習來探討汽車租賃公司與運輸網路公司之定價策略★ 以賽局理論分析入口網站應對內容農場之策略
★ 以賽局理論分析傳統車輛租賃業與運輸共 享平台之合作商業模式★ 基於多重代理人機制之現金回饋平台定價策略分析
★ 分析應用程式平台之隱私保護設定:硬體營收與廣告分潤★ 企業在產品差異化市場下之區塊鏈導入決策
★ 強化學習研究垂直差異化之訂價策略:以ChatGPT為例
檔案 [Endnote RIS 格式]    [Bibtex 格式]    [相關文章]   [文章引用]   [完整記錄]   [館藏目錄]   至系統瀏覽論文 (2029-6-30以後開放)
摘要(中) 隨著網路逐漸普及,人們已習慣於搜尋引擎上尋找資料,而隨著生成式人工智慧風潮興起,相較於傳統提供網頁自行瀏覽,生成式人工智慧直接統整出搜尋結果,節省了瀏覽網頁的時間,並給出準確的回應,使用人工智慧搜尋引擎也成為人們的選擇。基於傳統搜尋引擎長時間在市場上的獨佔,如何面臨AI搜尋引擎的競爭會是一個重要的課題。
使用賽局理論,可以探討在市場競爭下所有個體的策略與均衡。因此,本論文使用賽局模型分析兩間廠商在產品差異化的狀況下,討論不同的市場環境會對平台的決策造成甚麼影響,探討如何決定付費價格與廣告數量,讓平台保有市場競爭力,制定最佳策略並獲取最大利潤。本論文會使用納許Q學習的方式,模擬現實世界的決策方式,利用高效率與更貼近真實世界狀況的模擬,觀察真實市場中可能的決策變化趨勢,找出賽局模型中的最佳均衡解,提供搜尋引擎廠商、廣告商和用戶在制定商業模式和投資決策時的重要參考依據。
摘要(英) With the widespread use of the internet, people have become accustomed to searching for information on search engines. With the rise of generative artificial intelligence, compared to traditional methods of browsing web pages, generative AI directly synthesizes search results, saving time and providing accurate responses. Using AI search engines has also become a popular choice. Given the long-term dominance of traditional search engines in the market, how to face the competition from AI search engines is an important issue. Using game theory, one can explore the strategies and equilibria of all participants in market competition. Therefore, this paper uses a game model to analyze the situation of product differentiation between two firms, discussing how different market environments affect the platform′s decisions. It explores how to set the pricing and the number of ads to maintain market competitiveness, formulate optimal strategies, and maximize profit. This paper will use the Nash Q-learning method to simulate real-world decision-making, leveraging high efficiency and simulations that are closer to real-world conditions. It observes potential decision-making trends in the real market, identifies the optimal equilibrium solution in the game model, and provides important reference points for search engine vendors, advertisers, and users when formulating business models and investment decisions.
關鍵字(中) ★ 賽局模型
★ 強化學習
★ 納許Q學習
★ 搜尋引擎
★ 人工智慧
關鍵字(英)
論文目次 目錄
摘要.................................................................................................................................. i
Abstract............................................................................................................................ii
誌謝.................................................................................................................................iii
圖目錄 ............................................................................................................................. v
1. 緒論.......................................................................................................................... 1
1.1 研究背景.............................................................................................................. 1
1.2 研究動機........................................................................................................ 1
1.3 研究目的........................................................................................................ 2
2. 文獻回顧................................................................................................................... 3
2.1 賽局理論與搜尋引擎............................................................................................ 3
2.2 多代理系統 .......................................................................................................... 4
3. 賽局模型................................................................................................................... 6
3.1 賽局模型設置....................................................................................................... 6
4. 納許 Q 學習設置....................................................................................................... 9
5. 結果分析................................................................................................................. 11
6. 結論........................................................................................................................ 17
7. 參考文獻................................................................................................................. 18
參考文獻 [1] 「A Brief Overview of ChatGPT: The History, Status Quo and Potential Future Development」. 引見於: 2024年3月6日. [線上]. 載於: https://www.ieee-jas.net/en/article/doi/10.1109/JAS.2023.123618
[2] K. Eliaz及R. Spiegler, 作者, 「A Simple Model of Search Engine Pricing」, Econ. J., 卷 121, 期 556, 頁 F329–F339, 11月 2011, doi: 10.1111/j.1468-0297.2011.02467.x.
[3] R. P. dos Santos, 作者, 「Enhancing Physics Learning with ChatGPT, Bing Chat, and Bard as Agents-to-Think-With: A Comparative Case Study」. arXiv, 2023年6月1日. doi: 10.48550/arXiv.2306.00724.
[4] L. Waltman及U. Kaymak, 作者, 「Q-learning agents in a Cournot oligopoly mode」, J. Econ. Dyn. Control, 卷 32, 期 10, 頁 3275–3293, 10月 2008, doi: 10.1016/j.jedc.2008.01.003.
[5] J. Hu及M. P. Wellman, 作者, 「Nash q-learning for general-sum stochastic games」, J. Mach. Learn. Res., 卷 4, 期 null, 頁 1039–1069, 12月 2003.
[6] H. Hotelling, 作者, 「Stability in Competition」, Econ. J., 卷 39, 期 153, 頁 41–57, 1929, doi: 10.2307/2224214.
[7] L. Liu, L. Feng, B. Xu及W. Deng, 作者, 「Operation strategies for an omni-channel supply chain: Who is better off taking on the online channel and offline service?」, Electron. Commer. Res. Appl., 卷 39, 頁 100918, 1月 2020, doi: 10.1016/j.elerap.2019.100918.
[8] L. Chen及R. C. King, 作者, 「To Lend is to Own: A Game Theoretic Analysis of the E-Book Lending Market」, Int. J. Electron. Commer., 卷 21, 期 3, 頁 334–362, 7月 2017, doi: 10.1080/10864415.2016.1319214.
[9] R.-B. Esteves及J. Resende, 作者, 「Personalized pricing and advertising: Who are the winners?」, Int. J. Ind. Organ., 卷 63, 頁 239–282, 3月 2019, doi: 10.1016/j.ijindorg.2018.11.003.
[10] W. Amaldoss, J. Du及W. Shin, 作者, 「Pricing Strategy of Competing Media Platforms」, Mark. Sci., 8月 2023, doi: 10.1287/mksc.2021.0092.
[11] E. Tse及W. Zhao, 作者, 「Competition In Search Engine Market」, J. Bus. Strateg., 卷 28, 期 2, Art. 期 2, 2011, doi: 10.54155/jbs.28.2.123-150.
[12] N. Y. Motlagh, M. Khajavi, A. Sharifi及M. Ahmadi, 作者, 「The Impact of Artificial Intelligence on the Evolution of Digital Education: A Comparative Study of OpenAI Text Generation Tools including ChatGPT, Bing Chat, Bard, and Ernie」. arXiv, 2023年9月5日. doi: 10.48550/arXiv.2309.02029.
[13] R. S. Sutton及A. G. Barto, 作者, 「Reinforcement Learning: An Introduction」.
[14] C. J. C. H. Watkins及P. Dayan, 作者, 「Q-learning」, Mach. Learn., 卷 8, 期 3, 頁 279–292, 5月 1992, doi: 10.1007/BF00992698.
[15] J. Vainer及J. Kukacka, 作者, 「Nash Q-learning agents in Hotelling’s model: Reestablishing equilibrium」, Commun. Nonlinear Sci. Numer. Simul., 卷 99, 頁 105805, 8月 2021, doi: 10.1016/j.cnsns.2021.105805.
[16] A. Kastius及R. Schlosser, 作者, 「Dynamic pricing under competition using reinforcement learning」, J. Revenue Pricing Manag., 卷 21, 期 1, 頁 50–63, 2月 2022, doi: 10.1057/s41272-021-00285-3.
[17] M. L. Littman, 作者, 「Friend-or-Foe Q-learning in General-Sum Games」, 收入 Proceedings of the Eighteenth International Conference on Machine Learning, 收入 ICML ’01. San Francisco, CA, USA: Morgan Kaufmann Publishers Inc., 6月 2001, 頁 322–328.
指導教授 張李治華(Jhang Li,Jhih-Hua) 審核日期 2024-7-15
推文 facebook   plurk   twitter   funp   google   live   udn   HD   myshare   reddit   netvibes   friend   youpush   delicious   baidu   
網路書籤 Google bookmarks   del.icio.us   hemidemi   myshare   

若有論文相關問題,請聯絡國立中央大學圖書館推廣服務組 TEL:(03)422-7151轉57407,或E-mail聯絡  - 隱私權政策聲明