The proposed project utilizes advances in key classic computer science problems such as search tree restructuring to implement an optimized auction system for acquisition/sale of media advertising time. The proposed system would move away from the current manually negotiated approach to an expressive auction approach in which buyers can express rich preferences and constraints through which they can more effectively conduct their campaigns. The optimization software of the PI currently underlies the nationwide kidney exchange. This effort will scale and adapt the PI's prior work in expressive auctions for transportation and medical markets to match the richness of the TV market. Based on requirements gathering to be undertaken, a prototype application for the TV ad market will be completed.
The proposed work has the potential to reap significant savings though efficiencies and added derived value obtained for the expressive auction approach. This is anticipated by the PI to be anywhere from 13-41% of the current $75B market. Moreover, implementation of these approaches has the potential to create fairer and open markets while consumers benefit from viewing of ads relevant to their interests and needs. The approach has broader application in other areas such as electricity and pollution rights/credits as well as other media markets. The effort has a relationship within an established innovation ecosystem at CMU and a new small company will be launched.