DescriptionThe Internet Economy includes various online markets with billions of transactions. In this dissertation, we study pricing-related problems in both advertising markets and labor markets.
In an online advertising market, the advertiser pays for showing the ads to his target users.
Discovered by previous works, the prices of showing ads to users with different attributes vary a lot. Motivated by this observation, we develop two targeting algorithms to help an advertiser reach more target users when he has a budget. The first algorithm with provable guarantee is for the case when all the user information is completely revealed, and the second one is for the case when the user information is partially present. We also crawled LinkedIn and Facebook price to verify our algorithm. We further point out that these two algorithms are feasible only if the pricing has arbitrage. As price arbitrage may hurt the market revenue, we introduce arbitrage-free pricing for such markets, and finally propose three arbitrage-free pricing algorithms with provable revenue guarantee for the market.
In a labor market with multiple workers and tasks, a worker possibly has several skills and a task requires a worker with a certain skill. To fairly match workers and tasks, we introduce the stable pricing mechanism by extending stable matching. We propose three truthful stable pricing mechanisms with revenue guarantee to ensure the fairness for both workers and tasks.