optimal mechanisms and the polyamorous agent

a found poem

in many situations
agents
have strict restrictions
on their action spaces

(due to behavioral
or regulatory reasons)

for any multi-linear
social-choice
the optimal mechanism
with K
incurs an unexpected loss

of O

when compared to
optimal mechanisms
with unrestricted
action spaces


Found in « Implementation with a Bounded Action Space » by Liad Blumrosen & Michael Feldman in The Proceedings of the 7th ACM Conference on Electronic Commerce.

(You think that's suggestive -- you should read the paper on Learning from Revealed Preferences. Or maybe I just want the reading to more interesting than it is -- we've just entered hour three at O'Hare, which might have something to do with it.)