You are cordially invited to our Monday Lecture & Colloquium
on November 19th at 14:15 h & 16:00 h at TU Berlin. Location: Technische Universität Berlin Straße des 17. Juni 136 10623 Berlin Time: Monday, November 19th - 14:15 h Lecture: Jörg Rambau (Universität Bayreuth) Title: Optimal Diplomacy Picture yourself in a committee numerically evaluating a
scientific proposal that you find worth funding: A rating of "0"
means "easily achievable but not at all innovative", whereas "1"
means "very innovative but totally unachievable". In both cases,
funding is not recommended. In contrast, "1/2" means "innovative
and achievable", in other words: worth funding. All intermediate
values are possible. Any rating that is closer to "1/2" than to
"1/4" and "3/4" is considered a vote for funding. The proposal
passes if 50% of the members support funding, i.e., rate the
proposal between "1/2 - 1/8" and "1/2 + 1/8". Now, there are 10
meetings ahead of you. You have an idea how the opinions of the
committee members develop. How should your statements look like
in the meetings one through ten if you want to have eventually
as many supporters of the proposal as possible? This is an
instance of the "Optimal Diplomacy Problem" (ODP), introduced by
Hegselmann, König, Kurz, Niemann, and Rambau in 2010, published
in 2015. How do opinions interact? How is the dynamics of
opinions modeled mathematically? What does it mean to "influence
others" in this dynamical system? How difficult is it to find
optimal diplomacies? How can one compute or at least narrow down
optimal diplomacies? What happens if not all informations about
committee members are known to the diplomat? In this talk we
will discuss our findings, techniques, and open questions based
on the arguably most influential model: the Bounded-Confidence
model by Hegselmann and Krause. We draw on joint work with
Andreas Deuerling, Rainer Hegselmann, Stefan König, Julia
Kinkel, Sascha Kurz, and Christoph Niemann. Room MA 316 - Third Floor [British Reading] Time: Monday, November 19th - 16:00 h Colloquium: Jean-Philippe Labbé (Freie Universität
Berlin) Title: (At least) three hard problems behind the
multiassociahedron Abstract: |