Stochastic Models for Budget Optimization in Search-Based Advertising

Internet search companies sell advertisement slots based on users’ search queries via an auction. Advertisers have to solve a complex optimization problem of how to place bids on the keywords of their interest so that they can maximize their return (the number of user clicks on their ads) for a given budget. This is the budget optimization problem. In this paper, we model budget optimization as it arises in Internet search companies and formulate stochastic versions of the problem. The premise is that Internet search companies can predict probability distributions associated with queries in the future. We identify three natural stochastic models. In the spirit of other stochastic optimization problems, two questions arise. • (Evaluation Problem) Given a bid solution, can we evaluate the expected value of the objective function under different stochastic models? • (Optimization Problem) Can we determine a bid solution that maximizes the objective function in expectation under di...

• Type: Conference
• Year: 2007
• Where: WINE
• Authors: S. Muthukrishnan, Martin Pál, Zoya Svitkina
• Keywords: Economy, Internet Search Companies, Optimization Problem, Stochastic Models, WINE 2007