论文信息 - Markov Decision Process Based Adaptive Web Advertisements Scheduling

Markov Decision Process Based Adaptive Web Advertisements Scheduling

We study web advertisements scheduling problem by fully considering the interaction of web users and web advertisements publishing system. We construct a Markov Decision Process (MDP) based web advertisements scheduling model and schedule advertisements publishing during the whole process of web surfing by the users, thus we make maximal use of personal behavior characteristics of every web user in the scheduling model. We also track the user habit with reinforcement learning, solve the MDP model by TD(λ) algorithm combing the function approximator, and obtain adaptive online scheduling policies for web advertisements publishing.

[1] Kazuo Miyashita. Learning scheduling control knowledge through reinforcements , 2000 .

[2] Naoki Abe,et al. Improvements to the Linear Programming Based Scheduling of Web Advertisements , 2005, Electron. Commer. Res..

[3] Naoki Abe,et al. Unintrusive Customization Techniques for Web Advertising , 1999, Comput. Networks.

[4] David Maxwell Chickering,et al. Targeted Advertising on the Web with Inventory Management , 2003, Interfaces.

[5] Richard S. Sutton,et al. Reinforcement Learning: An Introduction , 1998, IEEE Trans. Neural Networks.

[6] Xinyi Peng. The Effective Genetic Clustering Algorithm and The Optimization of The Web Advertising Investment , 2006, 2006 International Conference on Computational Intelligence and Security.

[7] Naoki Abe,et al. Learning to Optimally Schedule Internet Banner Advertisements , 1999, ICML.