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。 ;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.
