An artificial society-based simulation framework for sponsored search auctions

Sponsored search advertising is an important channel for advertisers to reach potential consumers. Due to the nature of ascending auctions, complicated searching behaviors, and interactions among participants, it is difficult for advertisers and search providers to manipulate sponsored search auctions. In this research, we propose an artificial society-based simulation framework to facilitate modeling advertising objects, search behaviors, and underlying processes to support various related decisions in sponsored search auctions. The framework takes the stochastic cellular automata to mimic local interactions among stakeholders and capture search marketing dynamics. It also provides multiple advertisement retrieval, ranking, and pricing algorithms and a set of flexible agent interaction rules to support practical scenarios. We implement a Search Auction Experimental Platform (SAEP) to validate the proposed simulation framework. Preliminary experiments show that this framework helps to understand effects of competition levels and heterogeneous advertising strategies on sponsored search markets.