This week, Dr Talal Rahwan, a research fellow at the University of Southampton, dropped by our department at Masdar Institute to talk about the work he is doing on algorithms for reward maximization and reward distribution in agent coalition formation.
He gave us an overview of the improved dynamic programming algorithm (and one of its hybrid versions) for reward-maximal coalition formation, briefly presented his work on using domain-specific properties of the reward functions in reward maximization, and described the reward distribution problem and some of its algorithmic solutions. He used graphical representations a lot, making much of his material accessible to even those without formal acquaintance with the theory at a graduate level (I think).
Finally, he described applications in power generation management, marketing referral networks, network analysis for node targeting, etc. It was a very interesting and well-delivered presentation, and it was nice to see some solid algorithm work after quite a while.